Bee Colony Capstone - regression models.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "f4219b70",
   "metadata": {},
   "source": [
    "# Bee Colony Capstone - modeling step, using 3 regression models"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c55956c7",
   "metadata": {},
   "source": [
    "## Import modules"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "f2f3c58b",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import time\n",
    "import statsmodels.formula.api as smf\n",
    "\n",
    "from sklearn import linear_model\n",
    "from sklearn import model_selection\n",
    "from sklearn import pipeline\n",
    "from sklearn import svm\n",
    "from sklearn import ensemble\n",
    "from sklearn import metrics\n",
    "from patsy import dmatrices\n",
    "from patsy import dmatrix"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cc4b1933",
   "metadata": {},
   "source": [
    "## Import data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "19ed4c83",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = pd.read_csv('X_train.csv')\n",
    "X_test = pd.read_csv('X_test.csv')\n",
    "y_train = pd.read_csv('y_train.csv')\n",
    "y_test = pd.read_csv('y_test.csv')\n",
    "\n",
    "y_train = y_train['LogFinalToStart'].ravel()\n",
    "y_test = y_test['LogFinalToStart'].ravel()\n",
    "\n",
    "# This will only be used for the statsmodel formula, prior to indexing removing the unmodified columns\n",
    "bee_colony_data = pd.read_csv('bees_post_eda.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9ccc69c7",
   "metadata": {},
   "source": [
    "## Design matrices specification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "fc62a003",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['region_west', 'region_midwest', 'region_northeast', 'region_south', 'varroa_mites', 'Pesticides', 'region_west:varroa_mites', 'region_west:Pesticides', 'region_midwest:varroa_mites', 'region_midwest:Pesticides', 'region_northeast:varroa_mites', 'region_northeast:Pesticides', 'region_south:varroa_mites', 'region_south:Pesticides']\n"
     ]
    }
   ],
   "source": [
    "# naive predictors\n",
    "X_train_null = np.ones(shape=y_train.shape[0]).reshape(-1, 1)\n",
    "\n",
    "# Rename the varroa mites column to make it easier to work with\n",
    "X_train.rename(columns = {'Varroa mites': 'varroa_mites'}, inplace=True)\n",
    "X_test.rename(columns = {'Varroa mites': 'varroa_mites'}, inplace=True)\n",
    "\n",
    "# Regions only\n",
    "X_train_regions = X_train[['region_midwest', 'region_northeast', 'region_south', 'region_west']]\n",
    "X_test_regions = X_test[['region_midwest', 'region_northeast', 'region_south', 'region_west']]\n",
    "\n",
    "# Subregions only - breaking the larger regions down further\n",
    "subregion_cols = [col for col in X_train.columns.values.tolist() if col.startswith(\"subregion_\")]\n",
    "X_train_subregions = X_train[subregion_cols]\n",
    "X_test_subregions = X_test[subregion_cols]\n",
    "\n",
    "# All pest and other causes of colony loss columns\n",
    "X_train_all_pests = X_train[['varroa_mites', 'Other pests', 'Diseases', 'Pesticides', 'Other', 'Unknown']]\n",
    "X_test_all_pests = X_test[['varroa_mites', 'Other pests', 'Diseases', 'Pesticides', 'Other', 'Unknown']]\n",
    "\n",
    "# Varroa mites column plus subregion columns\n",
    "# I tried to use the subregion_cols variable here, but I couldn't figure out how to make it work with the mite column\n",
    "X_train_mites_subregions = X_train[['varroa_mites', 'subregion_east north central', 'subregion_east south central',\n",
    "       'subregion_mid atlantic', 'subregion_mountain', 'subregion_new england',\n",
    "       'subregion_pacific', 'subregion_south atlantic',\n",
    "       'subregion_west north central', 'subregion_west south central']]\n",
    "X_test_mites_subregions = X_test[['varroa_mites', 'subregion_east north central', 'subregion_east south central',\n",
    "       'subregion_mid atlantic', 'subregion_mountain', 'subregion_new england',\n",
    "       'subregion_pacific', 'subregion_south atlantic',\n",
    "       'subregion_west north central', 'subregion_west south central']]\n",
    "\n",
    "# Varroa mites data with larger regions\n",
    "X_train_mites_regions = X_train[['varroa_mites', 'region_midwest', 'region_northeast', 'region_south', 'region_west']]\n",
    "X_test_mites_regions = X_test[['varroa_mites', 'region_midwest', 'region_northeast', 'region_south', 'region_west']]\n",
    "\n",
    "# Patsy matrices\n",
    "\n",
    "# Using a negative one to exclude the intercept\n",
    "patsy = \"(region_west + region_midwest + region_northeast + region_south) * (varroa_mites + Pesticides) - 1\"\n",
    "# When I did the encoding or the indexing in the preprocessing step, it stripped out the \n",
    "# plain region / state / subregion columns and replaced them with the dummy columns\n",
    "# So, I went ahead and added the regions together and multiplied them by the varroa mites and pesticides\n",
    "\n",
    "train_model_matrix = dmatrix(patsy, X_train)\n",
    "train_model_matrix_names = train_model_matrix.design_info.column_names\n",
    "\n",
    "test_model_matrix = dmatrix(patsy, X_test)\n",
    "test_model_matrix_names = test_model_matrix.design_info.column_names\n",
    "print(test_model_matrix_names)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6a50410",
   "metadata": {},
   "source": [
    "## Feature coefficients with statsmodels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "0a521398",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:           percent_lost   R-squared:                       0.332\n",
      "Model:                            OLS   Adj. R-squared:                  0.294\n",
      "Method:                 Least Squares   F-statistic:                     8.602\n",
      "Date:                Sat, 12 Mar 2022   Prob (F-statistic):           2.33e-46\n",
      "Time:                        18:56:29   Log-Likelihood:                -2807.2\n",
      "No. Observations:                 878   AIC:                             5712.\n",
      "Df Residuals:                     829   BIC:                             5946.\n",
      "Df Model:                          48                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "===================================================================================================\n",
      "                                      coef    std err          t      P>|t|      [0.025      0.975]\n",
      "---------------------------------------------------------------------------------------------------\n",
      "Intercept                         411.1521    222.179      1.851      0.065     -24.948     847.253\n",
      "region[T.northeast]                28.0471     16.668      1.683      0.093      -4.670      60.764\n",
      "region[T.south]                    39.3727     20.716      1.901      0.058      -1.289      80.035\n",
      "region[T.west]                     30.9362     17.706      1.747      0.081      -3.817      65.689\n",
      "subregion[T.east south central]    19.1043      7.839      2.437      0.015       3.717      34.492\n",
      "subregion[T.mid atlantic]          16.0905      8.840      1.820      0.069      -1.260      33.441\n",
      "subregion[T.mountain]              17.3550      9.158      1.895      0.058      -0.621      35.331\n",
      "subregion[T.new england]           11.9566      7.866      1.520      0.129      -3.484      27.397\n",
      "subregion[T.pacific]               13.5813      8.570      1.585      0.113      -3.240      30.402\n",
      "subregion[T.south atlantic]        11.1078      6.772      1.640      0.101      -2.184      24.399\n",
      "subregion[T.west north central]    44.5699     24.880      1.791      0.074      -4.266      93.406\n",
      "subregion[T.west south central]     9.1605      6.214      1.474      0.141      -3.037      21.358\n",
      "State[T.Arizona]                -3.015e-12   1.63e-12     -1.852      0.064   -6.21e-12    1.81e-13\n",
      "State[T.Arkansas]                   6.5936      1.962      3.361      0.001       2.743      10.444\n",
      "State[T.California]                 5.6711      2.482      2.285      0.023       0.800      10.542\n",
      "State[T.Colorado]                   3.8372      1.988      1.930      0.054      -0.065       7.739\n",
      "State[T.Connecticut]                5.5656      4.114      1.353      0.176      -2.510      13.641\n",
      "State[T.Florida]                    2.5011      1.471      1.700      0.089      -0.386       5.388\n",
      "State[T.Georgia]                    1.7505      1.478      1.185      0.236      -1.150       4.651\n",
      "State[T.Hawaii]                     1.8648      3.011      0.619      0.536      -4.046       7.776\n",
      "State[T.Idaho]                      1.9886      1.937      1.027      0.305      -1.814       5.791\n",
      "State[T.Illinois]                  55.4491     28.476      1.947      0.052      -0.444     111.343\n",
      "State[T.Indiana]                   53.5351     28.474      1.880      0.060      -2.355     109.425\n",
      "State[T.Iowa]                       6.2215      3.717      1.674      0.095      -1.075      13.518\n",
      "State[T.Kansas]                    11.7450      3.763      3.121      0.002       4.360      19.130\n",
      "State[T.Kentucky]                  -3.2913      1.889     -1.742      0.082      -6.999       0.416\n",
      "State[T.Louisiana]                 -1.3499      1.977     -0.683      0.495      -5.230       2.530\n",
      "State[T.Maine]                   2.383e-15    8.4e-16      2.835      0.005    7.33e-16    4.03e-15\n",
      "State[T.Maryland]                   0.6506      1.633      0.398      0.691      -2.555       3.857\n",
      "State[T.Massachusetts]          -3.762e-16    8.6e-16     -0.437      0.662   -2.06e-15    1.31e-15\n",
      "State[T.Michigan]                  51.6897     28.486      1.815      0.070      -4.224     107.604\n",
      "State[T.Minnesota]                  6.0684      3.758      1.615      0.107      -1.308      13.445\n",
      "State[T.Mississippi]               -5.3223      2.415     -2.204      0.028     -10.063      -0.582\n",
      "State[T.Missouri]                   7.5512      3.744      2.017      0.044       0.203      14.899\n",
      "State[T.Montana]                   -0.3284      2.058     -0.160      0.873      -4.367       3.710\n",
      "State[T.Nebraska]                   4.0028      3.775      1.060      0.289      -3.408      11.413\n",
      "State[T.New Jersey]                 1.7110      3.180      0.538      0.591      -4.531       7.953\n",
      "State[T.New Mexico]                 7.7079      2.354      3.274      0.001       3.087      12.328\n",
      "State[T.New York]                   6.7382      3.108      2.168      0.030       0.638      12.838\n",
      "State[T.North Carolina]             1.3140      1.513      0.868      0.385      -1.656       4.284\n",
      "State[T.North Dakota]               4.5490      3.790      1.200      0.230      -2.891      11.989\n",
      "State[T.Ohio]                      55.2255     28.464      1.940      0.053      -0.645     111.095\n",
      "State[T.Oklahoma]                   3.6237      2.027      1.788      0.074      -0.355       7.603\n",
      "State[T.Oregon]                     2.3513      2.490      0.944      0.345      -2.536       7.238\n",
      "State[T.Pennsylvania]               7.6412      3.122      2.448      0.015       1.513      13.769\n",
      "State[T.South Carolina]            -0.3793      1.522     -0.249      0.803      -3.367       2.608\n",
      "State[T.South Dakota]               4.4321      3.776      1.174      0.241      -2.979      11.843\n",
      "State[T.Tennessee]                 -2.8688      1.971     -1.456      0.146      -6.737       0.999\n",
      "State[T.Texas]                      0.2931      1.882      0.156      0.876      -3.401       3.987\n",
      "State[T.Utah]                       2.0100      1.985      1.013      0.311      -1.885       5.905\n",
      "State[T.Vermont]                    6.3910      4.241      1.507      0.132      -1.934      14.716\n",
      "State[T.Virginia]                   2.9237      1.602      1.826      0.068      -0.220       6.067\n",
      "State[T.Washington]                 3.6940      2.492      1.483      0.139      -1.197       8.585\n",
      "State[T.West Virginia]              2.3472      1.527      1.537      0.125      -0.651       5.345\n",
      "State[T.Wisconsin]                 52.3267     28.440      1.840      0.066      -3.497     108.150\n",
      "State[T.Wyoming]                    2.1398      2.075      1.031      0.303      -1.933       6.213\n",
      "quarter[T.Q2]                      -7.4459      0.545    -13.670      0.000      -8.515      -6.377\n",
      "quarter[T.Q3]                      -4.4341      0.654     -6.785      0.000      -5.717      -3.151\n",
      "quarter[T.Q4]                      -3.7414      0.648     -5.770      0.000      -5.014      -2.469\n",
      "year                               -0.2222      0.124     -1.788      0.074      -0.466       0.022\n",
      "varroa_mites                        0.0324      0.015      2.126      0.034       0.002       0.062\n",
      "Pesticides                          0.0521      0.029      1.804      0.072      -0.005       0.109\n",
      "Diseases                           -0.0188      0.038     -0.492      0.623      -0.094       0.056\n",
      "==============================================================================\n",
      "Omnibus:                      405.086   Durbin-Watson:                   2.131\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             3834.160\n",
      "Skew:                           1.860   Prob(JB):                         0.00\n",
      "Kurtosis:                      12.538   Cond. No.                     1.05e+18\n",
      "==============================================================================\n",
      "\n",
      "Notes:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The smallest eigenvalue is 3.23e-27. This might indicate that there are\n",
      "strong multicollinearity problems or that the design matrix is singular.\n"
     ]
    }
   ],
   "source": [
    "bee_colony_data.rename(columns = {'Varroa mites': 'varroa_mites', 'Percent lost': 'percent_lost'}, inplace=True)\n",
    "\n",
    "mod = smf.ols(formula='percent_lost ~ region + subregion + State + year + quarter + varroa_mites + Pesticides + Diseases', data=bee_colony_data)\n",
    "res = mod.fit()\n",
    "print(res.summary())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a5dd5109",
   "metadata": {},
   "source": [
    "## Set up cross-validator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "b97e6765",
   "metadata": {},
   "outputs": [],
   "source": [
    "cross_validator = model_selection.KFold(n_splits=5, shuffle=True, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "71409eeb",
   "metadata": {},
   "source": [
    "## Linear ridge regression"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "30916a21",
   "metadata": {},
   "source": [
    "#### LRR  - create model object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "5008076f",
   "metadata": {},
   "outputs": [],
   "source": [
    "lr_baseline = linear_model.Ridge(fit_intercept=False, max_iter=500, random_state=42)\n",
    "\n",
    "lr_w_features = linear_model.Ridge(fit_intercept=True, max_iter=500, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a0fe08f9",
   "metadata": {},
   "source": [
    "#### LRR - create params object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "4da54d43",
   "metadata": {},
   "outputs": [],
   "source": [
    "lr_params = {\"model__alpha\": np.logspace(start=-10, stop=2, num=40)}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "64e731de",
   "metadata": {},
   "source": [
    "#### LRR - create pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "3c57ef76",
   "metadata": {},
   "outputs": [],
   "source": [
    "lr_pipeline = pipeline.Pipeline(steps=[(\"model\", lr_baseline)])\n",
    "\n",
    "lr_pipeline_w_features = pipeline.Pipeline(steps=[(\"model\", lr_w_features)])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "69f7f4db",
   "metadata": {},
   "source": [
    "#### LRR - create baseline grid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "c55d127c",
   "metadata": {},
   "outputs": [],
   "source": [
    "lr_baseline_grid = model_selection.GridSearchCV(\n",
    "                        estimator=lr_pipeline,\n",
    "                        param_grid=lr_params,\n",
    "                        refit=True,\n",
    "                        scoring='neg_mean_absolute_error',\n",
    "                        cv=cross_validator,\n",
    "                        verbose=1\n",
    "                    )\n",
    "\n",
    "lr_grid_w_features = model_selection.GridSearchCV(\n",
    "                        estimator=lr_pipeline_w_features,\n",
    "                        param_grid=lr_params,\n",
    "                        refit=True,\n",
    "                        scoring='neg_mean_absolute_error',\n",
    "                        cv=cross_validator,\n",
    "                        verbose=1\n",
    "                    )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "95800bc0",
   "metadata": {},
   "source": [
    "### Linear Ridge regression - naive model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b07b7652",
   "metadata": {},
   "source": [
    "#### LRR: naive - fit model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "67a07d0b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 40 candidates, totalling 200 fits\n",
      "Elapsed time: 0.32 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "lr_base_mod = lr_baseline_grid.fit(X_train_null, y_train)\n",
    "end = time.time()\n",
    "\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9fb3d6a7",
   "metadata": {},
   "source": [
    "#### LRR: naive - CV score distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "ad816530",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.01 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f948a072d60>"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEICAYAAACwDehOAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAAAcJ0lEQVR4nO3de5xVdb3/8ddbRuWiJiAZAgrlJS8PTBsVlRTDjoIaRl7IS2SanZOVdjE1u2iJmfXzp6fS5GjlLS+hJhpmhlKnNAwvoYAmeWMCAQkVUEn0c/5Y31kuhrlsYK/ZM8z7+XjMY9Z9fb5r79nvvb5rz9qKCMzMzAA2qnUBZmbWcTgUzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws51DoQiR9XdJVta5jbUg6T9L1FS47TdIpZddUiWLdkraVtFxSt3bY73OSDi57P63sPyRtX6v9r4taH7OOxqHQiaQn70JJvQrTTpE0rZL1I+LCiOgQL5pdSUS8EBGbRcRbrS0naYSkhvaqy6w5DoXOpw44vdZFdCWS6mpdg72jPc64ujKHQufzA+CrkrZsbqakyyTNk/SqpIclfagwr9il8VtJn2+y7t8kjU3D75d0r6R/SXpK0jEtFZS6bS6Q9EDqJrlTUl9JN6Q6/ippcGH5/dK0V9Lv/Qrzhkj6g6Rlku4Ftmqyr2FpPy+nekdUfuhW2855kiZJujnt6xFJuxfmPyfpLEkzgRWS6lrbd2t1SxqculXq0ngfST+XNF/SUkm/Tmd/dwPbpGO4XNI2kjaSdLakf0haIukWSX0K2z5R0vNp3rmttHeYpBeLL6iSPpbah6S9Jc1Ij9dCSZe0sq3PSJqbnhuTJW3TZJHRkp6R9JKkH0jaKK23fTpGr6R5Nxe22eLzTdIvJF0haYqkFcA5bbSlKsesy4oI/3SSH+A54GDgNuCCNO0UYFphmROAvmRnFF8BXgS6p3nnAden4U8Cfy6stwvwMrAp0AuYB5yUtrMn8BKwawt1TQPmAu8D3gXMBv6eaq0DrgV+npbtAywFTkzzPpHG+6b5DwKXpDoOAJYVah4ALAFGk72h+Uga71eo45Q0vG1qz7Yt1Hwe8CZwFLAx8FXgWWDjwrF+DBgE9Khg363VPRgIoC6N/wa4Geid9n1gmj4CaGhS5xnAX4CBadtXAjcWHrPlaX+bpv2vAg5uoc3/AD5SGP8VcHah/hPT8GbAsBa28eH0XNgz7fNHwB8L8wO4Pz3O26bnQeNjciNwbjp+3YHhaXqrzzfgF8ArwP6FdVtrS9WOWVf8qXkB/lmLB+udUNgt/ZH0o0koNLPOUmD3NHxe4YVqc2AFsF0anwD8LA0fC/xvk+1cCXy7hX1MA84tjP8/4O7C+BHAY2n4ROChJus/CHwqvYisAnoV5v2yUPNZwHVN1r0HGF+o45QKj+V5wF8K4xsBC4APFY71pwvzW9x3BXUPJoUC0B94G+jdTE0jWDMU5gAjC+P9ycKsDvgWcFNhXi/g3y29wAEXFB7jpo//H4Hzga3aOG5XAxcXxjdL9QxO4wEcWpj/OWBqGr4WmAgMbLLNVp9vZKFw7Vq0pWrHrCv+uPuoE4qIJ4C7gLObzpP0FUlz0in6y2Tv3LdqulxELCN7xzouTRoH3JCGtwP2Sd0kL6ftHA+8p5WyFhaGX29mfLM0vA3wfJN1nyd7J74NsDQiVjSZ12g74OgmdQ0n+6NfF/MaByLibaAh1bDG/Db23VbdRYOAf0XE0gpr3A64vbDPOcBbwNZpv8U2rCA7e2nJL4GxkjYFxgKPRERjnScDOwJPpi69w1vYxmqPX0QsT/scUFimeNye551j+jVAwEOSZkn6dKGNbT3fittsqy3VPGZdji+gdV7fBh4he1cOgLLrB2cBI4FZEfG2pKVkf4jNuRH4tqQ/knWR3J+mzwP+EBEfKaHu+WR/tEXbAr8le6feW1KvwgvstmTvPhvrui4iPlOlWgY1DqR+74GpvkbFWwi3uG9J27VRd9E8oI+kLSPi5SbzWlr+0xHx52b2uwDYuTDek6zrsFkRMVvS88Ao4DiyF9bGeU8Dn0jHYSwwSVLfJkEHTR6/dC2kL/DPwjKDgFlpeNu0DhHxIvCZtN5w4PfpuVfJ8221Y9NaW6jiMeuKfKbQSUXEXLJ+6S8WJm9O1o2xGKiT9C1gi1Y2M4XsD/w7wM3p3TJkZyE7pgtyG6efvSTt3OKWKjclbfs4ZRdvjyXr570rvdObAZwvaZP0wnFEYd3rgSMkHSKpm6Tuyj7GOXAda/mgpLHKLgCfAawk64tuTov7rqDuXEQsILugfLmk3unYHpBmLwT6SnpXYZWfAhNS8CCpn6Qxad4k4HBJwyVtQvY4tvU3/Uuy58wBZP3wpO2eIKlfeg68nCY39xHaXwInSfpAepd+ITA9Ip4rLHNmatsgsk/K3Zz2cXThsVpK9kL/Fuv+fGu2LVT/mHUpPhid23fI+kQb3UP2gvN3stP2N1jztDsXESvJLlofzOrvGpcB/0HWpTSf7GL198kuzK2XiFgCHE52EXwJWZfC4RHxUlrkOGAf4F9kZ0PXFtadB4wBvk4WfPOAM2nmeax3/mFs21bKuYOsP7vxwvfYiHizhbrb2neLdTfjRLI+7ieBRWSBREQ8SXb29kzq+tgGuAyYDPxO0jKy0NonLT8LOI3ssVuQ2tHW/zncSHbt4r7CMQc4FJglaXna57iIeKOZ4zAV+CZwa9rn+3inC7LRHcDDZBfqf0N2HQJgL2B62sdk4PSIeHY9nm8ttaXax6xLUbrYYtalSDoP2D4iTqh1LWYdic8UzMws51AwM7Ocu4/MzCznMwUzM8t16v9T2GqrrWLw4MG1LsOMp5Y8BcBOfXeqcSVmbXv44Ydfioh+zc3r1KEwePBgZsyYUesyzBjxixEATPvUtJrWYVaJ9I9/zXL3kZmZ5RwKZmaWcyiYmVmuU19TMLMN15tvvklDQwNvvLHG3TasQt27d2fgwIFsvPHGFa/jUDCzDqmhoYHNN9+cwYMHI7V0o19rSUSwZMkSGhoaGDJkSMXrufvIzDqkN954g759+zoQ1pEk+vbtu9ZnWg4FM+uwHAjrZ12On0PBzMxyDgUzs7Xw3HPPsdtuu9W6jDWMGDGiKv/M26UvNA8YtC3zG1r8Dhqzyn0q+6WT3N1RLXfffTcrVjT9NtDamz9/Pq+//npV76awatUq6urW7uV44002YfehQ6tWQ6MuHQrzG+Zx7JUP1LoM2wDct+Q0AD585U9qXMmGo2efHvTebvtal8HVV/yISTdeB8Axx4/nI6MOh426ceEPLmH24zMZ/L7t+eGPJ9KjZ08u/u63mHrPFOq61TF8xIc55/wLWfLSYr515hnM/2f2BW/f+O5FfHCffbns4gtZtHABDS+8QJ++fXnhuWf53qWXs+P7s28hPe7IUXz9/At57/Y78p2vf5Wn5sxm1apVfPHMc/jIqMNZ8NRjjBs3jtmzZ7Pzzjvz+uuvV6W9XToUzKxzuOBPZzH7pZlV3eYuWw3lG8O/3+oyT/ztUW696Xpuvft+IoKPjzqIvfcbzjNzn+Z7//8nfHCffTn79P/ihp//D0cddyL3TrmT3z3wCJJ49ZWXs9q/cRYnffY06oftx/yGeZx07JHc8+eH0/Yf4+Y7f0f3Hj342U9/zJQ7bmPH95/LooUvsujFBey2+x78cMJ5DBt+IBdddgWvvvIyYw8Zwf4HHMStt95Kz549mTlzJjNnzmTPPfesynHxNQUzsxbMmP4g/zHqCHr26kWvzTbjkMM+yoy/PED/AQP54D77AjDmqHHMmP4gm22+BZtu2p1zvnQa99x1B9179ATgz3+8n/PP+SpHHLQfnz3xWJYvW8by5csAGHnIaLr36AHA6DFjufvO2wGYcsdtjProxwD407T7uPJHl3DEQftx/JGjWblyJfP/OY9HH32UE07Ivk126NChDK1SV5LPFMysw2vrHX1ZWvoSsqYf9ZREXV0dt94zjQf/dxp33X4r1/1sItff9hvi7bf51ZSp+Yt/Uc+ePfPh9/Tfht69+/DkrCf4za9v5YIf/ndew09+dj3v3X7HNuuoBp8pmJm1YK999+feu+/i9dde47UVK/jdlDvzbqBH/jodgDtv/xUf3GdfVixfzvJXX2XEwYfwjQsuYs4TjwMwfMRIrrv6ynybsx9vuRvssCOPYuKPL2XZslfZaZddAfjQQSO59qor84Ca9fjfANhjjz244YYbAHjiiSeYObM63WsOBTOzFuw29AOMHXc8Yw8dwcdHHcQxx4/nXe/akvftuBO33/xLDjtwGK8sXcrxnzqFFSuW85kTjuawA4dx3JGjOPc73wPgmxMu5vG/PcphBw7jkOH13HjN1S3ub9QRR/KbX09i9EfH5tM+/+WzWPXmmxw2YhijDtibSy/6LgAf//jHWb58OUOHDuXiiy9m7733rkqbO/V3NNfX18f6fCxMkj99ZFWRf/qorz99VC2f+0APBgyp/aePOqqlzz9JfX19m8vNmTOHnXfeebVpkh6OiGZX9pmCmZnlHApmZpZzKJhZhxS0/Okfq8y6HD+Hgpl1SItfe5uVy15xMKyjxu9T6N69+1qt5/9TMLMOacozK4FF9Ov5Er6j1Jpe+9dLzJkzp9VlGr95bW04FMysQ3ptFUz6+8pal9Fh3fzZUaWcRbn7yMzMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs1ypoSDpS5JmSXpC0o2SukvqI+leSU+n370Ly58jaa6kpyQdUmZtZma2ptJCQdIA4ItAfUTsBnQDxgFnA1MjYgdgahpH0i5p/q7AocDlkrqVVZ+Zma2p7O6jOqCHpDqgJzAfGANck+ZfAxyZhscAN0XEyoh4FpgLVOerhMzMrCKlhUJE/BP4IfACsAB4JSJ+B2wdEQvSMguAd6dVBgDzCptoSNNWI+lUSTMkzVi8eHFZ5ZuZdUlldh/1Jnv3PwTYBugl6YTWVmlm2hp3e4qIiRFRHxH1/fr1q06xZmYGlNt9dDDwbEQsjog3gduA/YCFkvoDpN+L0vINwKDC+gPJupvMzKydlBkKLwDDJPWUJGAkMAeYDIxPy4wH7kjDk4FxkjaVNATYAXioxPrMzKyJ0r5PISKmS5oEPAKsAh4FJgKbAbdIOpksOI5Oy8+SdAswOy1/WkS8VVZ9Zma2plK/ZCcivg18u8nklWRnDc0tPwGYUGZNZmbWMv9Hs5mZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5UoNBUlbSpok6UlJcyTtK6mPpHslPZ1+9y4sf46kuZKeknRImbWZmdmayj5TuAz4bUS8H9gdmAOcDUyNiB2AqWkcSbsA44BdgUOByyV1K7k+MzMrKC0UJG0BHABcDRAR/46Il4ExwDVpsWuAI9PwGOCmiFgZEc8Cc4G9y6rPzMzWVOaZwnuBxcDPJT0q6SpJvYCtI2IBQPr97rT8AGBeYf2GNG01kk6VNEPSjMWLF5dYvplZ11NmKNQBewJXRMQewApSV1EL1My0WGNCxMSIqI+I+n79+lWnUjMzA8oNhQagISKmp/FJZCGxUFJ/gPR7UWH5QYX1BwLzS6zPzMyaKC0UIuJFYJ6kndKkkcBsYDIwPk0bD9yRhicD4yRtKmkIsAPwUFn1mZnZmupK3v4XgBskbQI8A5xEFkS3SDoZeAE4GiAiZkm6hSw4VgGnRcRbJddnZmYFpYZCRDwG1Dcza2QLy08AJpRZk5mZtcz/0WxmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZrmKQkHS/pVMMzOzzq3SM4UfVTjNzMw6sVbvfSRpX2A/oJ+kLxdmbQH4qzLNzDYwbd0QbxNgs7Tc5oXprwJHlVWUmZnVRquhEBF/AP4g6RcR8Xw71WRmZjVS6a2zN5U0ERhcXCciPlxGUWZmVhuVhsKvgJ8CVwH+4hszsw1UpaGwKiKuKLUSMzOruUo/knqnpM9J6i+pT+NPqZWZmVm7q/RMYXz6fWZhWgDvrW45ZmZWSxWFQkQMKbsQMzOrvYpCQdInm5seEddWtxwzM6ulSruP9ioMdwdGAo8ADgUzsw1Ipd1HXyiOS3oXcF0pFZmZWc2s662zXwN2qGYhZmZWe5VeU7iT7NNGkN0Ib2fglrKKMjOz2qj0msIPC8OrgOcjoqGEeszMrIYq6j5KN8Z7kuxOqb2Bf5dZlJmZ1Ual37x2DPAQcDRwDDBdkm+dbWa2gam0++hcYK+IWAQgqR/we2BSWYWZmVn7q/TTRxs1BkKyZC3WNTOzTqLSM4XfSroHuDGNHwtMKackMzOrlba+o3l7YOuIOFPSWGA4IOBB4IZ2qM/MzNpRW11AlwLLACLitoj4ckR8iews4dJySzMzs/bWVigMjoiZTSdGxAyyr+Y0M7MNSFuh0L2VeT2qWYiZmdVeW6HwV0mfaTpR0snAw+WUZGZmtdLWp4/OAG6XdDzvhEA9sAnwsUp2IKkbMAP4Z0Qcnr7G82ay7qfngGMiYmla9hzgZOAt4IsRcc/aNMbMzNZPq2cKEbEwIvYDzid7AX8OOD8i9o2IFyvcx+nAnML42cDUiNgBmJrGkbQLMA7YFTgUuDwFipmZtZNK7310f0T8KP3cV+nGJQ0EDgOuKkweA1yThq8BjixMvykiVkbEs8BcYO9K92VmZuuv7P9KvhT4GvB2YdrWEbEAIP1+d5o+AJhXWK4hTVuNpFMlzZA0Y/HixaUUbWbWVZUWCpIOBxZFRKUXpNXMtFhjQsTEiKiPiPp+/fqtV41mZra6Sm9zsS72Bz4qaTTZR1u3kHQ9sFBS/4hYIKk/0HhPpQZgUGH9gcD8EuszM7MmSjtTiIhzImJgRAwmu4B8X0ScAEwGxqfFxgN3pOHJwDhJm0oaQvZ1nw+VVZ+Zma2pzDOFllwE3JL+1+EFsu9oICJmSboFmE327W6nRcRbNajPzKzLapdQiIhpwLQ0vAQY2cJyE4AJ7VGTmZmtyd+JYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWa60UJA0SNL9kuZImiXp9DS9j6R7JT2dfvcurHOOpLmSnpJ0SFm1mZlZ88o8U1gFfCUidgaGAadJ2gU4G5gaETsAU9M4ad44YFfgUOBySd1KrM/MzJooLRQiYkFEPJKGlwFzgAHAGOCatNg1wJFpeAxwU0SsjIhngbnA3mXVZ2Zma2qXawqSBgN7ANOBrSNiAWTBAbw7LTYAmFdYrSFNa7qtUyXNkDRj8eLFpdZtZtbVlB4KkjYDbgXOiIhXW1u0mWmxxoSIiRFRHxH1/fr1q1aZZmZGyaEgaWOyQLghIm5LkxdK6p/m9wcWpekNwKDC6gOB+WXWZ2Zmqyvz00cCrgbmRMQlhVmTgfFpeDxwR2H6OEmbShoC7AA8VFZ9Zma2proSt70/cCLwuKTH0rSvAxcBt0g6GXgBOBogImZJugWYTfbJpdMi4q0S6zMzsyZKC4WI+BPNXycAGNnCOhOACWXVZGZmrfN/NJuZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWc6hYGZmOYeCmZnlHApmZpZzKJiZWa7DhYKkQyU9JWmupLNrXY+ZWVfSoUJBUjfgJ8AoYBfgE5J2qW1VZmZdR4cKBWBvYG5EPBMR/wZuAsbUuCYzsy6jrtYFNDEAmFcYbwD2KS4g6VTg1DS6XNJT67PDmz+7X+PgVsBL67OtTshtrrKb2a/thdqfH+cNlKTGwbVt73YtzehooaBmpsVqIxETgYlV37E0IyLqq73djsxt7hrc5g1fNdvb0bqPGoBBhfGBwPwa1WJm1uV0tFD4K7CDpCGSNgHGAZNrXJOZWZfRobqPImKVpM8D9wDdgJ9FxKx22n3Vu6Q6Abe5a3CbN3xVa68iou2lzMysS+ho3UdmZlZDDgUzM8t12VCQ1EfSvZKeTr97t7DclpImSXpS0hxJ+7Z3rdVSaZvTst0kPSrprvassdoqabOkQZLuT4/vLEmn16LW9dHW7WGU+e80f6akPWtRZzVV0ObjU1tnSnpA0u61qLOaKr0NkKS9JL0l6ai13UeXDQXgbGBqROwATE3jzbkM+G1EvB/YHZjTTvWVodI2A5xO525ro0ravAr4SkTsDAwDTutMt1ep8PYwo4Ad0s+pwBXtWmSVVdjmZ4EDI2Io8F06+cXnSm8DlJb7PtkHdtZaVw6FMcA1afga4MimC0jaAjgAuBogIv4dES+3U31laLPNAJIGAocBV7VPWaVqs80RsSAiHknDy8jCcEB7FVgFldweZgxwbWT+AmwpqX97F1pFbbY5Ih6IiKVp9C9k//fUmVV6G6AvALcCi9ZlJ105FLaOiAWQvSgA725mmfcCi4Gfp66UqyT1as8iq6ySNgNcCnwNeLud6ipTpW0GQNJgYA9gevmlVU1zt4dpGmqVLNOZrG17TgbuLrWi8rXZZkkDgI8BP13XnXSo/1OoNkm/B97TzKxzK9xEHbAn8IWImC7pMrLuh29WqcSqW982SzocWBQRD0saUcXSSlOFx7lxO5uRvcM6IyJerUZt7aTN28NUuExnUnF7JB1EFgrDS62ofJW0+VLgrIh4q3BfpLWyQYdCRBzc0jxJCyX1j4gF6TS6uVOtBqAhIhrfNU6i9X74mqtCm/cHPippNNAd2ELS9RFxQkklr7cqtBlJG5MFwg0RcVtJpZalktvDbGi3kKmoPZKGknWDjoqIJe1UW1kqaXM9cFMKhK2A0ZJWRcSvK91JV+4+mgyMT8PjgTuaLhARLwLzJO2UJo0EZrdPeaWopM3nRMTAiBhMdpuR+zpyIFSgzTYr+wu6GpgTEZe0Y23VUsntYSYDn0yfQhoGvNLYrdZJtdlmSdsCtwEnRsTfa1BjtbXZ5ogYEhGD09/vJOBzaxMIjRvpkj9AX7JPozydfvdJ07cBphSW+wAwA5gJ/BroXevay25zYfkRwF21rrvsNpN1K0R6jB9LP6NrXftatnM08HfgH8C5adp/Av+ZhkX2yZV/AI8D9bWuuR3afBWwtPCYzqh1zWW3ucmyvwCOWtt9+DYXZmaW68rdR2Zm1oRDwczMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPL/R+lZz2g4O87ogAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr_base_best_estimator = lr_base_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-validator score\n",
    "start = time.time()\n",
    "lr_cv_scores = model_selection.cross_val_score(\n",
    "                    lr_base_best_estimator,\n",
    "                    X_train_null,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "# Predict y on training data\n",
    "y_pred_train_naive = lr_base_best_estimator.predict(X_train_null)\n",
    "sns.histplot(y_pred_train_naive).set(title='Naive model: predicted vs observed')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "21e97dc7",
   "metadata": {},
   "source": [
    "#### LRR: naive - train / test comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "2a6cca58",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test_null = np.ones(shape=y_test.shape[0]).reshape(-1, 1)\n",
    "\n",
    "# Predict y on test data\n",
    "y_pred_test_naive = lr_base_best_estimator.predict(X_test_null)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "deb9b711",
   "metadata": {},
   "source": [
    "#### LRR: naive - calculate metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "2edcb8ef",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.066\n",
      "Train MAE:  0.050\n",
      "Test MAE:  0.050\n",
      "Train R-squared: -0.014\n",
      "Test R-squared: -0.020\n",
      "Train RMSE:  0.103\n",
      "Test RMSE:  0.084\n"
     ]
    }
   ],
   "source": [
    "print(f\"Mean train CV score: {np.mean(lr_cv_scores): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=y_pred_train_naive): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=y_pred_test_naive): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=y_pred_train_naive): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=y_pred_test_naive): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=y_pred_train_naive, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=y_pred_test_naive, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "177189b2",
   "metadata": {},
   "source": [
    "### Linear Ridge regression - with feature data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9726ee7c",
   "metadata": {},
   "source": [
    "#### LRR: with features - fit model - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "0d0fcb74",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 40 candidates, totalling 200 fits\n",
      "Elapsed time: 0.66 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "lr_regions_mod = lr_grid_w_features.fit(X_train_regions, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d7433484",
   "metadata": {},
   "source": [
    "#### LRR: with features - CV score distribution - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "d8204019",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.02 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f9481283b50>"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEICAYAAACwDehOAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAAAhNElEQVR4nO3de7xVdZ3/8ddbQblohYAOcgkrLMAYMsIKTUZNyUq0XyrOWHgbzNFKx2lSZyyd3zDjON6mHFNKUkvFY95ofl1UzIzJVDBEAflJI8IRhBOGQl4S/Mwf68tye9jnnM05e+19Lu/n47Efe+3v+q61Pt+1196ftb5r7b0UEZiZmQHsVO8AzMys83BSMDOznJOCmZnlnBTMzCznpGBmZjknBTMzyzkpdCKSbpD0z2n4IEnLa7TckPS+WiyrzLJHpuX3qsfy26ue6ywtv+Lto8bb0kWSfliLZdWapAskfa/ecRTNSWEHSVop6VVJmyWtk/R9SbtVezkR8auIeH8F8ZwkaX61l2/FqcYXZ6Xbx47WraXSnaCuICL+JSJOq3ccRXNSaJ/PRsRuwP7AR4B/bF6hq+35dmfKdJltvavF21X4M1kZb3gdEBHPAz8F9oO8S+FMSc8Az6Syz0haJGmjpF9LGrdtekkfkvS4pE2SbgP6lIybLKmx5PVwSXdKapK0QdLVkkYD1wIfS0cuG1PdXSVdJmlVOpq5VlLfknl9TdJaSWskndJS+yRNk7SgWdk5kuam4SMlLU3xPy/p71qYz06S/lHSc5LWS7pJ0jubVTslxbNW0rkl006UtEDSy6ktV5SM+2hapxslPSFpcsm4ByXNlPTfwCvABW20pSrrrC2SpgAXAMen9+yJFuJ9j6STJS1L6/d/JJ1eMp/m28dKSX8nabGklyTdJqnPjtZN4/++pK2nqZWuMkn7SPplivE+YFCz8bdLeiEt5yFJY1P5DOCvgL9P6+HHqfw8Sb9L81sq6ZgOrOvJkholfV3SC8D307a4bRkbJDVI2qNkmi+m7XSDpAvTujosjXvbEZ6koyQtSdvfg8o+j5W8H4Mk/Vea7kVJv1Jn2gmICD924AGsBA5Lw8OBJcD/Ta8DuA/YA+hLdiSxHjgA2BmYnqbfFdgFeA44B+gNfB54A/jnNK/JQGMa3hl4ArgS6E+WPA5M404C5jeL8Spgbopjd+DHwL+mcVOAdWSJrD9wS4r7fWXa2g/YBIwqKXsMmJaG1wIHpeEBwP4trLNTgBXAe4DdgDuBH6RxI9Pyb03xfBBoKlnHDwNfSMO7AR9Nw0OBDcCRZDs3n0yvB6fxDwKrgLFAL+CdbbSlKuuswm3oIuCHzcqax9sb+DTwXkDAwWTJYv/m20fJdvkosHdqwzLgS+2oOwV4IcXRD/hBa21N788VZNv0J9I6/mHJ+FPS+tw1reNFJeNuIG3vJWXHprh2Ao4H/ggMaed6ngxsAf4tLb8vcDbwG2BYKrsOuDXVHwNsBg4k+3xeRvaZ3LYt5u8bsG+K7ZPpvfp7sm18lwrW8b+S7cz1To+DANX7uy1fb/UOoKs90pu9GdhI9qV+DdA3jQvgkJK63yEljJKy5ekD/glgTenGAPya8knhY2RflL3KxHMSJUmB7Avkj8B7S8o+BjybhmcDl5SM27eND/0PgW+k4VHpQ98vvV4FnA68o411Ng/4m5LX708ftl68lRQ+UDL+UuD6NPwQcDEwqNk8v05KLCVlPwemp+EHgX+qpC3VXmcVbEMXUT4p/FMb090NfLX59lGyXZ7YbB1e2466s0nJML1+X0ttBUaQfen2Lym7pXnbSsa9K83rnen1DTRLCmWmWQRMbed6ngz8CehTUrYMOLTk9ZCSbfEbpASRxvVL05dLChcCDSV1dwKeByZXsI7/CbinvdtP0Y/Oc8jStRwdEe+KiHdHxN9ExKsl41aXDL8bODcdJm5U1r0znGzvYW/g+UhbSfJcC8sbDjwXEVsqiG0w2ca8sGSZP0vlpOWWxtjSMre5BTghDf8lcHdEvJJe/x+yPfXnUhfCx1qYx97NlvMc2Ydwr5Ky5jHtnYZPJfsSflrSY5I+k8rfDRzbbN0eSPYhLzfP1tpStXWm7EqfzemxpKV6LXhbvJI+Jek3qYthI9m6HlR2yswLJcOvkB1Z7Wjd5m1tvg5L7Q38ISL+WFKWrxtJO0u6JHXVvEz2RQmttCF13ywqeR/2a6l+yXreLGlEC7NsiojXSl6/G7irZP7LgK1k2+Lb2p62jQ0tzPdt23REvJmmHVpSp6V1/O9kRxX3pm7B81pYRl34xEv1lX7JrwZmRsTM5pUkHQwMlaSSxDAC+F2Zea4GRkjqVSYxNP+b298DrwJjIzvn0dxasiSzTUsfpm3uBQZJGk/2hXpOvuCIx4CpknoDZwENzea9zRqyD2PpMreQdckMS2XDgadLxq9Jy3gGOCH1uX4O+JGkgWTr5AcR8detxN583bTUlqqts4j4Fa1/GZeLa7tySbsCdwBfBO6JiDck3U12VFOktbz1nkD597O07gBJ/UsSwwjeasdfAlOBw8gSwjuBP/BWG962HiS9G/gucCjwcERslbSIFtoc2cUebWm+rlcDp0TEfzevKGkt2VHsttd9gYEtzHcNWVfntroiW1fltp+3BxSxCTiXbIdxLPALSY9FxLy2pq0FHykU67vAlyQdoEx/SZ+WtDtZX+wW4CuSekn6HDCxhfk8SvYBvCTNo4+kSWncOmCYpF0g32P5LnClpD0BJA2VdESq3wCcJGmMpH7AN1trQEpCPyLbu9mD7JwJknaR9FeS3hkRbwAvk+1xlXMrcI6yk5K7Af8C3NYswV0oqV/6kJwM3JaWc6KkwaldG1PdrWRdQZ+VdETaI+2j7MRi6RdaRW2p9jqrwDpgZBsnF3ch6/NuArZI+hRweAeXW4kG4GRJo1Nbv9FSxYh4DlgAXJy2hwOBz5ZU2R14nWxvux/Z+15qHdl5pm36k32JNwFIOpl0EUcVXQvMTAkISYMlTU3jfkS2TX08fZ4upuUk3AB8WtKhaafoXLK2/rqtAJRdfPK+lEi2fW5a+uzUnJNCgSJiAfDXwNVke0gryM4BEBF/ItvzPSmNO57sBGy5+Wwl+7C9j6wfvzHVB3iA7GT3C5J+n8q+npb1m3TYfj9pDygifkp2wu+BVOeBCppyC9ne3u3Nvsi/AKxMy/gScGIL088mO2H5EPAs8Brw5WZ1fpnimQdcFhH3pvIpwBJJm4H/IDsx/FpErCbbC72A7EtkNfA12t6mW2pLtddZa25PzxskPV6uQtqb/ArZl88fyPa653ZwuW1Kbf0W8Auytj6cRr3ewiR/SXYhxYtkyfKmknE3kXWxPA8sJTvBW+p6YEzqyrk7IpYCl6dlriPbE99uj76D/oNsPd4raVOK6QCAiFhCtl3OIdsJ20R2och2bY+I5WTb+7fJjjQ/S3ap+p8qiGEU2fa1mayt10TEgx1qVRXp7V3aZmZvUXaZ5VPArhWe0+o20lHtRrIr1p6tczg14yMFM3sbScek7qABZJdz/rinJARJn03dmP3JLkl9krdOkPcITgpm1tzpZF1yvyPr6z6jvuHU1FSyk8hryLp5pkUP605x95GZmeV8pGBmZrku/TuFQYMGxciRI+sdhplVwfIN2b97v39gp/tD125n4cKFv4+IweXGdemkMHLkSBYsWNB2RTPr9CbfMBmAB096sK5x9ASSWvxVvruPzMws56RgZmY5JwUzM8t16XMK5bzxxhs0Njby2muvtV3ZyurTpw/Dhg2jd+/e9Q7FzGqs2yWFxsZGdt99d0aOHEn2f1O2IyKCDRs20NjYyD777FPvcMysxrpd99Frr73GwIEDnRDaSRIDBw70kZZZD9XtkgLghNBBXn9mPVe3TApmZtY+Tgo1snLlSvbbr9r3C+m4yZMn+weAZpZzUujCtmzpEf9mbNYtDB0+AklVewwd3taddNunsKuPJPUhu9PWrmk5P4qIb0q6iOxuZE2p6gUR8ZM0zflkN2rfCnwlIn5eVHxFu+KKK5g9ezYAp512GkcffTRbtmxh+vTp/Pa3v2Xfffflpptuol+/fpx33nnMnTuXXr16cfjhh3PZZZfR1NTEl770JVatWgXAVVddxaRJk7joootYs2YNK1euZNCgQfzud79j9uzZjB07Fsj2/C+//HI+8IEP8OUvf5knn3ySLVu2cNFFFzF16lReffVVTj75ZJYuXcro0aN59dVX67aOzHqSNY2rOf66Nu/WWbHbTv941eZVqshLUl8HDomIzekepvMl/TSNuzIiLiutLGkMMA0YC+wN3C9p33QrynY5+2dns+iFRe2dvKzxfzaeq6Zc1WqdhQsX8v3vf59HHnmEiOCAAw7g4IMPZvny5Vx//fVMmjSJU045hWuuuYZTTjmFu+66i6effhpJbNy4EYCvfvWrnHPOORx44IGsWrWKI444gmXLluXznz9/Pn379uXKK6+koaGBiy++mLVr17JmzRo+/OEPc8EFF3DIIYcwe/ZsNm7cyMSJEznssMO47rrr6NevH4sXL2bx4sXsv//+VV0/Zta1FdZ9FJnN6WXv9Gjt5g1TgTkR8Xq69d0KWr6Rfac2f/58jjnmGPr3789uu+3G5z73OX71q18xfPhwJk2aBMCJJ57I/Pnzecc73kGfPn047bTTuPPOO+nXrx8A999/P2eddRbjx4/nqKOO4uWXX2bTpk0AHHXUUfTt2xeA4447jttvz27529DQwLHHHgvAvffeyyWXXML48eOZPHkyr732GqtWreKhhx7ixBOzWymPGzeOcePG1XTdmFnnVuiP1yTtDCwku+H8f0bEI5I+BZwl6YvAAuDciPgDMJS339i7MZU1n+cMYAbAiBGt96m1tUdflJZuXNT8Uk9J9OrVi0cffZR58+YxZ84crr76ah544AHefPNNHn744fzLv1T//v3z4aFDhzJw4EAWL17MbbfdxnXXXZfHcMcdd/D+92//N8S+5NTMWlLoieaI2BoR44FhwERJ+wHfAd4LjAfWApen6uW+qbb7do2IWRExISImDB5c9u/A6+4Tn/gEd999N6+88gp//OMfueuuuzjooINYtWoVDz/8MAC33norBx54IJs3b+all17iyCOP5KqrrmLRokUAHH744Vx99dX5PLeVlzNt2jQuvfRSXnrpJT74wQ8CcMQRR/Dtb387T1C//e1v89huvvlmAJ566ikWL15c7eabWRdWk6uPImIj8CAwJSLWpWTxJvBd3uoiagSGl0w2jOw+qV3O/vvvz0knncTEiRM54IADOO200xgwYACjR4/mxhtvZNy4cbz44oucccYZbNq0ic985jOMGzeOgw8+mCuvvBKAb33rWyxYsIBx48YxZswYrr322haX9/nPf545c+Zw3HHH5WUXXnghb7zxBuPGjWO//fbjwgsvBOCMM85g8+bNjBs3jksvvZSJE7tkD52ZFaSwezRLGgy8EREbJfUF7gX+DVgYEWtTnXOAAyJimqSxwC1kSWJvYB4wqrUTzRMmTIjm19gvW7aM0aNHF9KmnsTr0Wqtu99kR1LVrz5q7/e3pIURMaHcuCLPKQwBbkznFXYCGiLivyT9QNJ4sq6hlcDpABGxRFIDsBTYApzZkSuPzMxsxxWWFCJiMfChMuVfaGWamcDMomIyM7PWdctfNBfVJdZTeP2Z9VzdLin06dOHDRs2+IutnbbdT6FPnz71DsXM6qDb3WRn2LBhNDY20tTU1HZlK2vbndfMrOfpdkmhd+/evmOYmVk7dbvuIzMzaz8nBTMzyzkpmJlZzknBzMxyTgpmZpZzUjAzs5yTgpmZ5ZwUzMws56RgZmY5JwUzM8s5KZiZWc5JwczMck4KZt3M0OEjkFS1x9DhI+rdJKuhbvcvqWY93ZrG1VW/F7D1HD5SMDOznJOCmZnlCksKkvpIelTSE5KWSLo4le8h6T5Jz6TnASXTnC9phaTlko4oKjYzMyuvyCOF14FDIuLPgfHAFEkfBc4D5kXEKGBeeo2kMcA0YCwwBbhG0s4FxmdmZs0UlhQiszm97J0eAUwFbkzlNwJHp+GpwJyIeD0ingVWABOLis/MzLZX6DkFSTtLWgSsB+6LiEeAvSJiLUB63jNVHwqsLpm8MZU1n+cMSQskLWhqaioyfDOzHqfQpBARWyNiPDAMmChpv1aqq9wsysxzVkRMiIgJgwcPrlKkZmYGNbr6KCI2Ag+SnStYJ2kIQHpen6o1AsNLJhsGrKlFfGZmliny6qPBkt6VhvsChwFPA3OB6anadOCeNDwXmCZpV0n7AKOAR4uKz8zMtlfkL5qHADemK4h2Ahoi4r8kPQw0SDoVWAUcCxARSyQ1AEuBLcCZEbG1wPjMzKyZwpJCRCwGPlSmfANwaAvTzARmFhWTmZm1zr9oNjOznJOCmZnlnBTMzCznpGBmZjknBTMzyzkpmJlZzknBzMxyTgpmZpZzUjAzs5yTgpmZ5ZwUzMws56RgZmY5JwUzM8s5KZiZWc5JwczMck4KZmaWc1IwM7Ock4KZmeWcFMzMLOekYGZmucKSgqThkn4haZmkJZK+msovkvS8pEXpcWTJNOdLWiFpuaQjiorNzMzK61XgvLcA50bE45J2BxZKui+NuzIiLiutLGkMMA0YC+wN3C9p34jYWmCMZmZWorAjhYhYGxGPp+FNwDJgaCuTTAXmRMTrEfEssAKYWFR8Zma2vZqcU5A0EvgQ8EgqOkvSYkmzJQ1IZUOB1SWTNVImiUiaIWmBpAVNTU1Fhm1m1uMUnhQk7QbcAZwdES8D3wHeC4wH1gKXb6taZvLYriBiVkRMiIgJgwcPLiZoM7MeqtCkIKk3WUK4OSLuBIiIdRGxNSLeBL7LW11EjcDwksmHAWuKjM/MzN6uyKuPBFwPLIuIK0rKh5RUOwZ4Kg3PBaZJ2lXSPsAo4NGi4jMzs+0VefXRJOALwJOSFqWyC4ATJI0n6xpaCZwOEBFLJDUAS8muXDrTVx6ZmdVWYUkhIuZT/jzBT1qZZiYws6iYzMysdf5Fs5mZ5ZwUzMws56RgZmY5JwUzM8s5KZiZWc5JwczMck4KZmaWc1IwM7Ock4KZmeWcFMzMLOekYGat26kXkqr6GDp8RL1bZS0o8g/xzKw7eHMLx1/366rO8rbTP17V+Vn1+EjBzMxyTgpmZpZzUjAzs5yTgpmZ5ZwUzMws56RgZma5ipKCpEmVlJmZWddW6ZHCtyssMzOzLqzVH69J+hjwcWCwpL8tGfUOYOc2ph0O3AT8GfAmMCsi/kPSHsBtwEhgJXBcRPwhTXM+cCqwFfhKRPy8HW0yM7N2autIYRdgN7LksXvJ42Xg821MuwU4NyJGAx8FzpQ0BjgPmBcRo4B56TVp3DRgLDAFuEZSq4nHzMyqq9UjhYj4JfBLSTdExHM7MuOIWAusTcObJC0DhgJTgcmp2o3Ag8DXU/mciHgdeFbSCmAi8PCOLNfMzNqv0v8+2lXSLLIun3yaiDikkokljQQ+BDwC7JUSBhGxVtKeqdpQ4DclkzWmsubzmgHMABgxwn+qZWZWTZUmhduBa4HvkfX3V0zSbsAdwNkR8bKkFquWKYvtCiJmAbMAJkyYsN14MzNrv0qTwpaI+M6OzlxSb7KEcHNE3JmK10kako4ShgDrU3kjMLxk8mHAmh1dppmZtV+ll6T+WNLfSBoiaY9tj9YmUHZIcD2wLCKuKBk1F5iehqcD95SUT5O0q6R9gFHAoxW3xMzMOqzSI4VtX+JfKykL4D2tTDMJ+ALwpKRFqewC4BKgQdKpwCrgWICIWCKpAVhKduXSmRGxQ11VZmbWMRUlhYjYZ0dnHBHzKX+eAODQFqaZCczc0WWZmVl1VJQUJH2xXHlE3FTdcMzMrJ4q7T76SMlwH7I9/cfJfrFsZmbdRKXdR18ufS3pncAPConIzMzqpr1/nf0K2dVBZmbWjVR6TuHHvPVDsp2B0UBDUUGZmVl9VHpO4bKS4S3AcxHRWEA8ZmZWRxV1H6U/xnua7B9SBwB/KjIoMzOrj0rvvHYc2a+LjwWOAx6R1NZfZ5uZWRdTaffRPwAfiYj1AJIGA/cDPyoqMDMzq71Krz7aaVtCSDbswLRmZtZFVHqk8DNJPwduTa+PB35STEhmZlYvbd2j+X1kN8X5mqTPAQeS/Z/Rw8DNNYjPzMxqqK0uoKuATQARcWdE/G1EnEN2lHBVsaGZmVmttZUURkbE4uaFEbGA7NacZmbWjbSVFPq0Mq5vNQMxM7P6ayspPCbpr5sXphvkLCwmJDMzq5e2rj46G7hL0l/xVhKYAOwCHFNgXGZmVgetJoWIWAd8XNJfAPul4v8XEQ8UHpmZmdVcpfdT+AXwi4JjMTOzOvOvks3MLFdYUpA0W9J6SU+VlF0k6XlJi9LjyJJx50taIWm5pCOKisvMzFpW5JHCDcCUMuVXRsT49PgJgKQxwDRgbJrmGkk7FxibmZmVUVhSiIiHgBcrrD4VmBMRr0fEs8AKYGJRsZmZWXn1OKdwlqTFqXtpQCobCqwuqdOYyrYjaYakBZIWNDU1FR2rmVmPUuuk8B3gvcB4YC1weSpXmbpRpoyImBUREyJiwuDBgwsJ0sysp6ppUoiIdRGxNSLeBL7LW11EjcDwkqrDgDW1jM3MzGqcFCQNKXl5DLDtyqS5wDRJu0raBxhFdvtPMzOroUpvsrPDJN0KTAYGSWoEvglMljSerGtoJXA6QEQskdQALAW2AGdGxNaiYjMzs/IKSwoRcUKZ4utbqT8TmFlUPGZm1jb/otnMzHJOCmZmlnNSMDOznJOCmZnlnBTMzCznpGBmZjknBTMzyzkpmJlZzknBzMxyTgpmZpZzUjAzs5yTgpmZ5ZwUzMws56RgZmY5JwUzM8s5KZiZWc5JwczMck4KZmaWc1IwM7Ock4KZmeUKSwqSZktaL+mpkrI9JN0n6Zn0PKBk3PmSVkhaLumIouIyM7OWFXmkcAMwpVnZecC8iBgFzEuvkTQGmAaMTdNcI2nnAmMzM7MyCksKEfEQ8GKz4qnAjWn4RuDokvI5EfF6RDwLrAAmFhWbmZmVV+tzCntFxFqA9LxnKh8KrC6p15jKtiNphqQFkhY0NTUVGqyZWU/TWU40q0xZlKsYEbMiYkJETBg8eHDBYZmZ9Sy1TgrrJA0BSM/rU3kjMLyk3jBgTY1jMzPr8WqdFOYC09PwdOCekvJpknaVtA8wCni0xrGZmfV4vYqasaRbgcnAIEmNwDeBS4AGSacCq4BjASJiiaQGYCmwBTgzIrYWFZuZmZVXWFKIiBNaGHVoC/VnAjOLisfMzNrWWU40m5lZJ+CkYGZmOScFMzPLOSmYmVnOScHMzHJOCmZmlnNSMDOznJOCmZnlnBTMzCznpGBmZjknBTMzyzkpmJlZzknBzMxyTgpmZpZzUjAzs5yTgpmZ5ZwUzMws56RgZmY5JwUzM8s5KZiZWa5XPRYqaSWwCdgKbImICZL2AG4DRgIrgeMi4g/1iM/MrKeq55HCX0TE+IiYkF6fB8yLiFHAvPTazMxqqDN1H00FbkzDNwJH1y8UM7OeqV5JIYB7JS2UNCOV7RURawHS857lJpQ0Q9ICSQuamppqFK6ZWc9Ql3MKwKSIWCNpT+A+SU9XOmFEzAJmAUyYMCGKCtDMrCeqy5FCRKxJz+uBu4CJwDpJQwDS8/p6xGZm1pPVPClI6i9p923DwOHAU8BcYHqqNh24p9axmZn1dPXoPtoLuEvStuXfEhE/k/QY0CDpVGAVcGwdYjMz69FqnhQi4n+APy9TvgE4tNbxmJnZWzrTJalmZlZnTgpmZpZzUjAzs5yTgpmZ5ZwUzMws56RgZmY5JwUzM8s5KZiZWc5JwczMck4KZmaWc1IwM7Ncj04KQ4ePQFLVHkOHj6h3k8zMOqReN9npFNY0rub4635dtfnddvrHqzYvM7N66NFHCmZm9nZOCmZmlnNSMDOznJOCmZnlnBTMzCznpGBmZjknBTMzy3W6pCBpiqTlklZIOq/e8ZiZ9SSdKilI2hn4T+BTwBjgBElj6huVmVnP0amSAjARWBER/xMRfwLmAFPrHJOZWY+hiKh3DDlJnwemRMRp6fUXgAMi4qySOjOAGenl+4HlFc5+EPD7KobbVfTEdvfENoPb3ZN0tM3vjojB5UZ0tv8+Upmyt2WtiJgFzNrhGUsLImJCewPrqnpiu3tim8HtrncctVRkmztb91EjMLzk9TBgTZ1iMTPrcTpbUngMGCVpH0m7ANOAuXWOycysx+hU3UcRsUXSWcDPgZ2B2RGxpEqz3+Eup26iJ7a7J7YZ3O6epLA2d6oTzWZmVl+drfvIzMzqyEnBzMxy3SopSNpD0n2SnknPA8rUGS7pF5KWSVoi6as7Mn1nU2nMkmZLWi/pqWbl4yX9RtIiSQskTaxN5B3T0XancV9Of6myRNKlxUfdcdVodxr/d5JC0qBiI+64Kmzj/y7paUmLJd0l6V01CbyDqtDudn2fdaukAJwHzIuIUcC89Lq5LcC5ETEa+ChwZslfaVQyfWdTacw3AFPKlF8KXBwR44FvpNddQYfaLekvyH4tPy4ixgKXFRRntXX0/UbScOCTwKoiAixAR9t8H7BfRIwD/j9wfhFBFqCj7W7f91lEdJsH2a+bh6ThIcDyCqa5B/hke6ev92NHYgZGAk81K/s5cHwaPgG4pd5tqlG7G4DD6t2OWrc7lf8I+HNgJTCo3m2qRZtLxh8D3FzvNtWi3e39PutuRwp7RcRagPS8Z2uVJY0EPgQ80p7pO4mOxnw28O+SVpPtLXeVvaiOtntf4CBJj0j6paSPVD3CYnSo3ZKOAp6PiCeKCK4g1fxcngL8tCpRFa+j7W7X9J3qdwqVkHQ/8GdlRv3DDs5nN+AO4OyIeLkasRWlWm1uwRnAORFxh6TjgOuBw6ow3w4ruN29gAFkXYgfARokvSfSblU9FdVuSf3SPA7vyHyKUPB7vW0Z/0DWfXxztebZUbVo947qckkhIlr8wpK0TtKQiFgraQiwvoV6vckSws0RcWfJqIqmr7VqtLkV04FtJ9tvB77XzjCrruB2NwJ3piTwqKQ3yf5krKn9EVdHge1+L7AP8IQkyP5G5nFJEyPihQ4F3UEFv9dImg58Bji0MyT+bQpud7um727dR3PJvuRIz/c0r6Ds03A9sCwirtjR6Tuhjsa8Bjg4DR8CPFOluIrW0XbfTdZeJO0L7ELX+KfNdrc7Ip6MiD0jYmREjCRLjPvXOyFUoEPvtaQpwNeBoyLilSrHVqSObuPtm77eJ1OqfGJmINlZ9mfS8x6pfG/gJ2n4QLJ/Xl0MLEqPI1ubvjM/Kmlzen0rsBZ4g+zL4NSS9bEQeILs3MqH692mGrV7F+CHwFPA48Ah9W5TLdrdbF4r6Ronmjv6Xq8AVpd83q+td5tq1O52fZ/5by7MzCzX3bqPzMysA5wUzMws56RgZmY5JwUzM8s5KZiZWc5JwczMck4KZmaW+1/IYSBkorD8AwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr_regions_best_estimator = lr_regions_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "lr_cv_scores_regions = model_selection.cross_val_score(\n",
    "                    lr_regions_best_estimator,\n",
    "                    X_train_regions,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "# Predict y on training data\n",
    "y_train_lr_pred_regions = lr_regions_best_estimator.predict(X_train_regions)\n",
    "sns.histplot(y_train_lr_pred_regions).set(title='Predicted vs observed - training data - regions')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "168a89f6",
   "metadata": {},
   "source": [
    "#### LRR: with features - metrics - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "id": "c72d1e21",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.066\n",
      "Train MAE:  0.049\n",
      "Test MAE:  0.052\n",
      "Train R-squared:  0.036\n",
      "Test R-squared:  0.040\n",
      "Train RMSE:  0.100\n",
      "Test RMSE:  0.081\n"
     ]
    }
   ],
   "source": [
    "y_test_lr_pred_regions = lr_regions_best_estimator.predict(X_test_regions)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(lr_cv_scores_regions): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=y_train_lr_pred_regions): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=y_test_lr_pred_regions): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=y_train_lr_pred_regions): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=y_test_lr_pred_regions): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=y_train_lr_pred_regions, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=y_test_lr_pred_regions, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dbef8056",
   "metadata": {},
   "source": [
    "#### LRR: with features - plotting observed vs predicted - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "id": "823b3f83",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))\n",
    "\n",
    "ax1.scatter(y_test, y_test_lr_pred_regions, c=\"orange\")\n",
    "\n",
    "p1 = max(max(y_test_lr_pred_regions), max(y_test))\n",
    "p2 = min(min(y_test_lr_pred_regions), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts', ylabel='Predicted', title=\"Linear Ridge regression on regions - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(y_test_lr_pred_regions), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title=\"Linear Ridge regression on regions - natural scale\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8f753bca",
   "metadata": {},
   "source": [
    "#### Performance\n",
    "\n",
    "The linear ridge regression on just region data performs only slightly better than the naive model. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "edd50825",
   "metadata": {},
   "source": [
    "#### LRR: with features - fit model - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "id": "40f619a9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 40 candidates, totalling 200 fits\n",
      "Elapsed time: 0.73 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "lr_subregions = lr_grid_w_features.fit(X_train_subregions, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e8d70c37",
   "metadata": {},
   "source": [
    "#### LRR: with features - CV score distribution - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "id": "14861cbe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.03 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f9468ea8fa0>"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr_subregions_best_estimator = lr_subregions.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "lr_cv_scores_subregions = model_selection.cross_val_score(\n",
    "                    lr_subregions_best_estimator,\n",
    "                    X_train_subregions,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "lr_train_pred_subregions = lr_subregions_best_estimator.predict(X_train_subregions)\n",
    "sns.histplot(lr_train_pred_subregions).set(title='Predicted vs observed - training data - subregions')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "94e2d713",
   "metadata": {},
   "source": [
    "#### LRR: with features - metrics - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "id": "18111708",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.066\n",
      "Train MAE:  0.052\n",
      "Test MAE:  0.052\n",
      "Train R-squared:  0.057\n",
      "Test R-squared:  0.082\n",
      "Train RMSE:  0.099\n",
      "Test RMSE:  0.080\n"
     ]
    }
   ],
   "source": [
    "lr_subregions_test_pred = lr_subregions_best_estimator.predict(X_test_subregions)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(lr_cv_scores_subregions): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=lr_train_pred_subregions): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=lr_subregions_test_pred): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=lr_train_pred_subregions): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=lr_subregions_test_pred): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=lr_train_pred_subregions, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=lr_subregions_test_pred, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2579ae4d",
   "metadata": {},
   "source": [
    "#### LRR: with features - plotting observed vs predicted - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "id": "f35a2acb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0gAAAFNCAYAAAA6tU9BAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAABraklEQVR4nO3de9wUZf3/8deHG1BvxVMCiYqQmoWVioiHNE0xD8VX0wyR0igjxMrO+pMsy0w7Z6kopWkKHsJU8JAKHkhNOQgq4AnFIyR4SvTOAzef3x/XbOy99+7O7L2zO3vf9/v5eNyPvee6Zub6XNfM7uy1M3ONuTsiIiIiIiICPbIOQEREREREpFGogyQiIiIiIhJRB0lERERERCSiDpKIiIiIiEhEHSQREREREZGIOkgiIiIiIiIRdZDqwMz2NbPHs44jibhYzexSM/tZPWNqVGZ2i5kdn3UctWBmz5jZiDqVdZqZ/bkeZSVlZneZ2QlZxxHHzPY3sxeyjkMaU6Mfe+Le+/X8HGp0ZvammX0g6zjSZmaDzMzNrGedyuuyx+1aibbP9nUq60tmdk89yoqjDlKKSn2Yu/s/3X3HLGIqZGZnmNl70Yft62Z2n5ntlctvpFgbnbsf6u6XZR1HZ+fuP3f3hu+MiDSqznDsKUbv/eTcfSN3fzrrODq7rnbc1o8ItaMOUhdW5heZq919I2AL4E7gb/WLKpm0f00ys6Y01yfr1OuXPxHpHJJ8JjTS50aax4dGqldXpPZNl9qzNHWQ6qDwMpiox/89M3vYzP5jZleb2fp5+Z8xs4V5Z3g+lpd3qpk9ZWarzWyJmX02L+9LZnavmf3OzF4FzigXl7uvAaYAW5lZ3xKx7mpmD0blXQ2sn78OM/uBma0ws+VmdkL+qVgzW8/Mfm1mz5nZS2Z2oZltUKKN2sUet3xM2Zea2SQzu9nM3gI+aWYDzOxaM1tlZsvM7Jt56xpuZvPM7I2orN9G6eub2RVm9kq0PeaaWf8o73+XYZlZDzP7oZk9a2YrzeyvZrZJlJe7hOD4qC4vm9nEUtvFzDaJll8Vre+HZtYjr53uidrltageh5ZZ1ylm9mK0/R43swPz2udnefMVu1Rr92gfe83M/pLbR3PzRuv+N/CXqP65ffMVM7vGzDbPW/9xUV1eMbPTLe9XLwtnNa/Im/f/zGxx1N53mdmH8/JKvnfMbAszuzFa7lUz+2eu3apRbtvG1a3Iug6L2nR1tF2+l5d3uIX3/RtROx4SpY81s0ejZZ42s6+VibXkPi7dT+H7utz7J8qv2bEnep9Ps/B5+gbwpSLv/S/mvZcmFiy/gZldFn0ePWrh8z+/bon3fav8+BBX9jMWPg8fBt4ys55mtmfUhq+b2UNmtn9Bez0dteUyMxsTpW9vZndH2+ZlC8fc3DL5x7c0jxEftvA5+7qFz93/K2in883spijWB8xsuxLrKXesbPOZWLjdI1+2cCxfYWbfLZi3cL/ZxMwujuZ90cx+ZlEn18yazOw3UfstM7OvW94lfJbScdtKfGfoiKic8Wb2ZLTNzjczi/K2M7M7onZ92cymmNmmUd7lwEBghoWrgn5gRY7l1v54W9iew83sX9F2W2Fm55lZ74SxF92Xo7yv2rpj1xIzGxqll/wsKbL+D5nZ7RaO6Y+b2ecra90quLv+UvoDngFGFEnfH3ihYL45wABgc+BRYHyUNxRYCewBNAHHR/OvF+UfHS3XAxgFvAVsGeV9CVgDfAPoCWxQJJYzgCui/3sD5wAvAz0LY43ynwW+DfQCPge8B/wsyj8E+DewE9AMXA44sH2U/3tgelTHPsAM4OwSbdcu9nLLJyj7UuA/wMejtmoG5gM/iur1AeBp4OBo/n8BX4z+3wjYM/r/a1G5zdH22A3YOMq7Czgh+v/LwNJovRsBfwcuj/IGRbH9KarXzsA7wIdLtMVfgRuiOg8CngC+ktdO7wFfjeI5EVgOWJH17Ag8DwzIi2O7vPb5Wcw+ugjYJmr/e/O2+/7RtvoFsF5Up28B9wNbR2kXAVdG8w8B3gT2idr+11EdRhTZJz9I2KcPIuxzP4jatXeC987ZwIXRcr2AfYu1S8L3ctJtW7ZuRda7Atg3+n8zYGj0/3DC/noQYX/dCvhQlPdpYDvAgP2Alrzl/rfdouVK7uP667p/dJ5jz3vAEdE6Nih47+feS58gfIb8Nlpn7nPiHODu6H2zNfBwR/d9Kj8+lCw7r10XEj4vN4jev68Ah0XrPyia7gtsCLwB7BgtuyWwU/T/lcDEaJn1gX3yysg/vqV1jOhF+Gw7Lar3AcDqvNguBV4lfD71JPygelWJNi13rHyGvP2zYLsPiup2ZdQ2HwVW0fb4ULjfXE84xmwI9CPs01+L5h8PLIm202bAzGj9ue84d5HCcZsS3xk6+P514EZgU0KHZxVwSJS3PWH/WY+w/8wGfl/qvU/Be75wnhLtuRuwZ7SNBxE+F75VbN8rWG+5fflo4EVgd8Kxa3tg24SfJffkrf95YGwU21DC99Wd6vK5Wo9Custf4Y5aaoeN5vtC3vQvgQuj/ycBZxYs/ziwX4kyFwKH+7od67mYGM8A3gVeB1oJH9r7F4uVcKBq88EK3Me6L8qXkNfhid4AHr1atNNvl5e/F7CsRFxtYo9bvlzZ0fSlwF/z8vcobBvg/wF/if6fDfwE2KJgni9Hdf5YkZjvYt0H7SxgQl7ejoQPodwHjgNb5+XPAY4pss4mwofwkLy0rwF35bXT0ry85mjd7y+yru0JX3hGAL0K8i4lvoM0Pm/6MOCpvHnfBdbPy38UODBvesu8+v+IqLOUF/O7FO8gnQ5ckzdvD8KH7P4J3js/JXxpaPdB3oH3ctJtW7ZuRdb7XLQ9Ny5Ivwj4XcLYrgdOLvJ+LbuP66/r/tF5jj2zi6Tl3vs/Iu/LN+HLUf7nRJsOD3BCR/d9Kj8+lCw7r12/nDd9CtEX7by0Wwmdzg0Jx9+jKOhIEjo+k8k7VuTl5Y6taR4j9iX80NgjL+1K4Iy8dvpzXt5hwGMl2rTcsbLN/knxDtKHCvbLi4vtN0D/qP4b5KWNBu6M/r+DqLMUTY+gdAepw8dtSnxn6MhfVE5+Z/ga4NQS8x4BLCjTtvsT30GaHRPPt4DrCve9IvOV25dvJTpOJaj/Qtp+luQ6SKOAfxbMexHw42rbPMmfLrHLzr/z/m8h/AIBsC3w3ehU5+tm9jrhV6kB8L/LeRbm5X2EcC9RzvMJyr7G3TclfNAsIvx6UMwA4EWP9srIswX5+eXl/9+X6Fe5vFj/EaWXUsny5coulrYtMKCgXU8jtAHAVwhnLx6LLg34TJR+OeGNflV0+v+XZtarSFkDaNs2zxI+ZPvnpZXa5vm2YN2Zu/x1bVVsPe7eEv3bbl3uvpTwQXcGsNLMrjKzAUXKLCW//Z4l2gcjq9z97bzpbYHr8tr2UUIHvD8F2yqK+ZUSZbZpR3dfGy1btP60bcdfEX4NvC065X9qsQIsjJz1ZvR3YYk4SsZE221bSd0gHEgOA561cClNboCUbYCnSsR7qJndH11i8Hq0/BZFZo3bx0Ug22NPuXkK30tv0fa9VO4zvyP7fiXLd+R4c3TB+vYh/Er+FuGL33hghYXL1z4ULfcDwo+Dcyxc7vblIuWkdozI1Sv6nI1dF6WPW5D8WFlKueNNYdv2IrRdrm0vIpxJgmTbKqea43ap7wxtWBg1L3e8GVNsnnLlmFm/6Nj9YnRJ3BUU//yvRJs2MbMPWrg8/d9RGT9PUkbMvlzumBb3WZKzLbBHwftoDPD+RLWskjpIjed54Cx33zTvr9ndrzSzbQmne78OvC/q5CwifKDmePtVFufuLxN+eTrDzLYsMssKwv1J+esfWJC/dd70Nnn/vwz8l3AqNFePTTwMDlEypAqWL1d2sfU9Tzj7lN+ufdz9MAB3f9LdRxM+ZH8BTDOzDd39PXf/ibsPAfYGPgMcV6Ss5YQ3c85AwuUhL5WpbzEvE37BKlzXixWuBwB3n+ru+0Trc0LdIJyda86btdgHTn6bDiTU8X+rLpj3eeDQgvZd391fpGBbWbiP7H0lQm7TjtG+tw0J6u/uq939u+7+AWAk8B2L7rkqmO/nHkaE2sjdx8ettzAm2m7bSuqGu89198MJ+9n1hF8KIbRfu2v7zWw94FrCpXv9o/f8zbR9z+eU3cdFYtTj2FNunhXkfeaYWTNt30vlPvM7su8nPj7ElF1qfZcXrG9Ddz8HwN1vdfeDCGfaHyO0Le7+b3f/qrsPIBybL7D2wyuneYxYDmxjbe/V7NC6Yo6VaR5vniecQdoir203dvedovwk2yqnw8ftUt8Zisx3aN7xZkrceos4m1D/j7n7xsAXKP++a9PWFu7NKvxhunCZSYT9cIeojNMofoxpp9S+TOljWpLPkpzngbsL3kcbufuJSWKrljpI6etl4WbF3F+lI4T8CRhvZntYsKGZfdrM+hBOZzrh+lTMbCyh591h7v4Y4VefHxTJ/hfhw+KbFm46PZJwLXLONcBYCzd5NhMukcitd21Ul9+ZWb8o3q3M7OCEccUtX7LsEuYAb1i4kXYDCzdyfsTMdo/W/QUz6xuV+3q0TKuZfdLMPhp9yLxBODC1Fln/lcC3zWywmW1E+AXmag8DYSTm7q1R3c4ysz7Rh8l3CL8aVcTMdjSzA6Iv2W8TOpy52BcCh5nZ5mb2fsKZpkInmdnWFgZbOA24usg8ORdGMW8bld3XzA6P8qYBI81sbws3fv6E0h++1wCfNrMDo18fv0s4GN6XoL6fsXCTsxG2VSvFt1Wlym3bxHUzs95mNsbMNnH39/JiBLiYsD8faOHG4a2iX+J6E649XwWssXCz9adKxFl2H5cur1Mde4qYBnzGzPaJ3ks/pe13lGuA/2dmm5nZVoQvWDnV7vtxy5cru5grCJ8LB0frWt/CzfNbm1l/CwPRbEj4bHuT6HPAzI42s9yX+9cIbd7mMyzNYwTwAOEL9Q/MrJeFgSRGAldVuqKYY+VC4JiojGGE+5kLnW5mzWa2E+Gek6LHG3dfAdwG/MbMNo4+L7czs/2iWa4BTo4+QzclXO5YSoeP26W+M8Qt1wF9CPvI69G+9/2C/JcI91DlPAGsH713ewE/JBxD4sp4A3gzOu4k6oCU25eBPwPfM7Pdos+T7aN9tZLPkhuBD1oYvKVX9Le75Q3cVEvqIKXvZsIX0dzfGZUs7O7zCDdXnkf4gFxKuCYTd18C/IbQcXmJcDPjvSnE/CtgnEUdkbxY3gWOjMp/jXAq9e95+bcAfyAMFb40igvCGwXCB9NS4H4Lp21nEq7xTark8gnKbiM6qIwEdgGWEX6F+zOwSTTLIcBiM3sTOJdwnfHbhF+6phE+PB4l3Khb7EB0CeESg9nR+t8m3LDcEd8gHLSeBu4Bpkbrr9R6rBuE49+EX7pOi/IuBx4iXJt8G8UPRlOjvKejv3IPCD6XMKDGbWa2mjBgwx4A7r44qtNVhF/3VhPujWq3rdz9ccIvZH+M4h4JjIz2xTg7EPaRNwn7wwXufleC5eKU3LaV1C3yReCZaH8eT6gr7j6H8KXgd4Sbx+8m3NC6Gvgm4aD/GnAsoZ3bSbCPS9fWGY89+eUvBk4ifO6siGLIH43rp9H0MsL7fBrR+6zafT/B8iXLLrG+54HDCZ+3qwi/hH+f8J2rB+GHn+WEARD2AyZEi+4OPBAdh6YT7uFYVqSIVI4R0efq/wGHRnW+ADgu+uG0UuWOlacTzia8RvgRaWqR5e8m7HOzgF+7+21lyjqO8OPRkmid0whnMCB09G8jDKSxgPC+WEPxzks1x+1S3xnS9hPC4AT/AW4i7ztY5GzghxYuQfueu/+HsD/9mXAm8C3avo+K+R7h2LKa0H7lfgzNV3Jfdve/AWcRtvVqwhUTm1fyWRId/z4FHBOV8W/WDQ5Vc+Ztbi8R6bioV7+IMOpRRWdOOnPZUpnol7rXCafzix38O62uXDeRRmJmJxK+lO4XO3MXKlsqE511v9Ddt42dWSSPziBJVczss9GlQ5sRevYz6tVBybJsqYyZjYwun9iQcD/NI4SzV51eV66bSKMwsy3N7OPRJVU7En65vq6rly2ViS6RPMzCbQFbAT9G20o6QB0kqdbXCJcQPEU4hV2Xm+caoGypzOGEU+TLCZfCHeNd5/R1V66bSKPoTRitbDVhKOcbCJeEdfWypTJGuCztNcIldo8Sf4+ySDu6xE5ERERERCSiM0giIiIiIiIRdZBEREREREQilT4noVPYYostfNCgQVmHISLSrc2fP/9ldy98SKGg45SISCModZzqkh2kQYMGMW/evKzDEBHp1szs2axjaFQ6TomIZK/UcUqX2ImIiIiIiETUQRIREREREYmogyQiIiIiIhJRB0lERERERCSiDpKIiIiIiEhEHSQREREREZGIOkgiIiIiIiKRTDtIZnaImT1uZkvN7NQi+WZmf4jyHzazoVnEKSIiIiINas4EuLInTLXwOmdC+mUsmwLXD4KpPcLrsin1X0ca9YyLIUkZceuIy585Iqw/9zdzRGXL10FmD4o1sybgfOAg4AVgrplNd/clebMdCuwQ/e0BTIpeRURERKS7mzMBlk5aN+2t66aHX5BOGcumwJxx0NoSplueDdMAg8fUZx1p1DMuhiRlxK0jLn/mCFg5q21cK2eF9BEz02nrFGR5Bmk4sNTdn3b3d4GrgMML5jkc+KsH9wObmtmW9Q5URKQ7eecd+NWvwquISEN7anJl6R3x0MR1X9hzWltCer3WkUY942JIUkbcOuLyCztHObn0NNo6BVl2kLYCns+bfiFKq3QeAMxsnJnNM7N5q1atSjVQEZHu4tln4ROfgB/8AG66KetoRERieGtl6R3R8lxl6bVYRxr1jIshSRlx66i2nmm0dQqy7CBZkTTvwDwh0X2yuw9z92F9+/atOjgRke7mlltg6FB49FGYNg2OPDLriEREYlhTZekd0TywsvRarCONesbFkKSMuHVUW8802joFWXaQXgC2yZveGljegXlERKQKra1w+ulw2GGw9dYwfz4cdVTWUYmIJLDduMrSO2Lns6CpuW1aU3NIr9c60qhnXAxJyohbR1x+vwOLl5FLT6OtU5BlB2kusIOZDTaz3sAxwPSCeaYDx0Wj2e0J/MfdV9Q7UBGRrmrlSjj4YPjZz2DsWLj/fthhh6yjEhFJaPgFsP2J685yWFOYTmuABgiDAwyfDM3bAhZeh0+ubNCAateRRj3jYkhSRtw64vJHzGzfSep3YEhPo51SYu5Fr1irT+FmhwG/B5qAS9z9LDMbD+DuF5qZAecBhwAtwFh3nxe33mHDhvm8ebGziYh0a/fcA6NGwauvwvnnw5e/nO76zWy+uw9Ld61dg45TIiLZK3WcymyYbwB3vxm4uSDtwrz/HTip3nGJiHRl7vDb38Ipp8DgwXDzzbDzzllHJSIi0hgy7SCJiEh9vf56uJTu+uvDIAyXXAKbbJJ1VCIiIo0jy3uQRESkjhYuhGHD4MYbwxmkadPUOcoxs0PM7HEzW2pmpxbJ38zMrjOzh81sjpl9JErfxszuNLNHzWyxmZ1c/+hFRCRN6iCJiHRx7nDxxbDnnvD223DXXfDtb4MVe5BCN2RmTcD5wKHAEGC0mQ0pmO00YKG7fww4Djg3Sl8DfNfdPwzsCZxUZFkREelE1EESEenCWlrCJXUnnAD77gsPPggf/3jWUTWc4cBSd3/a3d8FrgIOL5hnCDALwN0fAwaZWX93X+HuD0bpq4FHKfFAcxER6RzUQRIR6aKeeCKcNfrrX+FHP4J//AP69cs6qoa0FfB83vQLtO/kPAQcCWBmw4FtCc/m+x8zGwTsCjxQq0BFRKT2NEiDiEgXNG1aGLa7d+8wSt0hh2QdUUMrdrFh4TMwzgHONbOFwCPAAsLldWEFZhsB1wLfcvc3ihZiNg4YBzBwYH2fCi8iIsnpDJKISBfy7rvwrW/B0UfDTjvBggXqHCXwArBN3vTWwPL8Gdz9DXcf6+67EO5B6gssAzCzXoTO0RR3/3upQtx9srsPc/dhffv2TbkKIiKSFnWQRES6iOefh/32g3PPhZNPhrvvhm22iV9OmAvsYGaDzaw3cAwwPX8GM9s0ygM4AZjt7m9EDzS/GHjU3X9b16hFRKQmdImdiEgXcOutMGYMvPMOXH01fP7zWUfUebj7GjP7OnAr0ARc4u6LzWx8lH8h8GHgr2bWCiwBvhIt/nHgi8Aj0eV3AKdFD0IXEZFOSB0kEZFOrLUVzjwTfvrTcEndtGmw445ZR9X5RB2amwvSLsz7/1/ADkWWu4fi9zCJiEgnpQ6SiEgntWpVOGt0++1w3HEwaRI0N2cdlYiISOeme5BERDqh++6DXXeF2bNh8mS49FJ1jkSkQS2bAtcPgqk9wuuyKfWPYeYImGrr/maOaJs/ZwJc2TPkXdkzTFeSD/H1jMufsVPbGGfsVFkdkpQRV48k26ratoqrZ5K2rjF1kEREOhF3+P3vw2AM660H//oXfPWrYLrIS0Qa0bIpMGcctDwLeHidM66+naSZI2DlrLZpK2et62DMmQBLJ4G3hmlvDdO5L+Zx+RBfz7j8GTvB6iVtY1y9ZF3nIa4OScqIq0eSbVVtW8XVM0lb14E6SCIincQbb4Thu7/9bfj0p2H+/HAWSUSkYT00EVpb2qa1toT0einsWBSmPzW5eH4uPS4f4usZl1/YacjJpcfVIUkZcfVIsq2qbau4eiZp6zpQB0lEpBN4+GEYNgyuvx5+9Su47jrYdNOsoxIRidHyXGXpWcidrSiVHpcP8fWsRzvElRFXjyQxptFW5VS7fErUQRIRaXCXXgp77AFvvgl33gnf+54uqRORTqJ5YGXpWbCm8ulx+RBfz3q0Q1wZcfVIEmMabVVOtcunRB0kEZEG9d//wgknwNixsNdesGAB7Ltv1lGJiFRg57OgqWAEmabmkF4v/Q4sn77duOL5ufS4fIivZ1x+nyHFy8ilx9UhSRlx9Uiyraptq7h6JmnrOlAHSUSkAS1dGjpFF18MEyeGobz79886KhGRCg0eA8MnQ/O2gIXX4ZNDer2MmNm+g9HvwJAOMPwC2P7EtmdBtj8xpCfJh/h6xuWPXNy+89BnSEhPUockZcTVI8m2qrat4uqZpK3rwNy9rgXWw7Bhw3zevHlZhyEi0iHXXQdf+hI0NcEVV8Bhh2UdUceY2Xx3H5Z1HI1IxykRkeyVOk7pDJKISIN47z347nfhyCNhxx3DJXWdtXMkIiLSWfXMOgAREYEXX4RRo+Dee+Gkk+A3vwnPORIREZH6UgdJRCRjM2fCscdCSwtMnQqjR2cdkYiISPelS+xERDKydi2ceSZ86lPQty/MnavOkYiISNZ0BklEJAMvvwxf/CL84x8wZgxcdBFsuGHWUYmIiIg6SCIidfbAA3D00fDSSzBpEnzta3rwq4iISKPQJXYiInXiDn/8Y3jYa1MT3HcfjB+vzpGIiEgjUQdJRKQOVq+GY46Bb34TDj4YHnwQdtst66hERESkkDpIIiI1tmgR7L47TJsG55wDN9wAm22WdVQiIiJSjO5BEhGpocsvD/cYbbwxzJoF+++fdUQiIiJSjs4giYjUwNtvh47RccfB8OGwYIE6RyIiIp2BOkgiIil7+mnYe2+YPBlOPTU8CHbLLbOOSkRERJJQB0lE6mPOBLiyJ0y18DpnQnrrXjYFrh8EU3uE12VTar9sifpMnw5Dh8KyZeH/s8+GnvkXM5crr1wblVuuVN7MEWFdub+ZI6qvt4iISBene5BEpPbmTIClk9ZNe+u66eEXVLfuZVNgzjhobQnTLc+GaYDBY2qzbJH6rHl8MhPPO4JfXv4phg4NAzIMHlxBeavuLd1GfT9eejkonrfo57B6SdvyV84KnaTtxna8zURERLo4c/esY0jdsGHDfN68eVmHISI5V/YMX/gLWROMXlPduq8fFL7gF2reFo54pjbLFtRnxWvvZ9Qfr+afj3+C8ePhd7+D9devsLz/vlC6jTbYuvRyUDyvnOZtO95mFTCz+e4+LLUVdiE6TomIZK/UcUqX2InIOrW67KrYF//C9HKXg0Hpy89aniu+7lx6uTrFLTtjp7YxzdipXdx3Lt6fXScuYP4zu3H5iV9g0iRYf0WJMsuVV66Nyi1XKq+cuHqLiIh0Y7rETkSCai5Vi2NNpc+OQOgMrZzVNi93OdiImeUv0Wtqhta32q+7qTm+Tr02h/deab9sr81DZ6jwErXVS/7XSVq71jhnxqmc/rcz+eCWTzDrtAPZaeslsOzQ0mWWi7X1v8Da9nn0gJ7NsObN9lk9NwyvxfLK6blh+fWJiIh0YzqDJCLBQxPXfanPaW0J6dXablz59MLOUU4u/anJxfOfmhx1LIpo/W98naxEvEb7zlHO6iW8+uZm/N9vpjPxmp/z+T2vYe6Zu4fOEZQvs1ysPTYontdjA1hTpFMFIb1UXjnl1iciItLN6QySiAS1vOwqNxDDU5PD2R9rCp2jpAM0JLlEr5218XV699Xi+aXSgblPDePoP/yN5a8N4LzjT2LCQRdg+R2tsmWWuudzLaxtKZHVUma5jt5Dmvb6REREug6dQRKRoHlgZemVGn5BGJDhWA+vlYxel7sUr1h6uby4OlVQZ3e44PYT2een9+Bu3PPjfTjpUwWdo7h1djTWjtS/nHLr66bM7BAze9zMlprZqUXyNzOz68zsYTObY2Yfycu7xMxWmtmi+kYtIiK1oA6SiAQ7nxXuhcnX1BzSa63fgeXTy12iVy4vrk7l8vsM+V/Sm29vyJjzp3DSpRcwYpf7eXDytxm+3dzi8ZZbZ0dj7Uj9mzYtnt5nSPwlj92MmTUB5wOHAkOA0WY2pGC204CF7v4x4Djg3Ly8S4FD6hCqiIjUgS6xE5EgNxDDQxPD5WDNA8OX83o8F2fEzPYDNfQ7MKRDskv0yuWVqlO5Og8eAzN2YsmjzufOncbjK3bkrC/8nlMv+xY9euxXPt5y6+xIrEmWK5ZXONBEnyEwcnGyNutehgNL3f1pADO7CjgcyL8RbQhwNoC7P2Zmg8ysv7u/5O6zzWxQvYMWEZHayOQ5SGa2OXA1MAh4Bvi8u79WZL5LgM8AK939I4X5pej5EiKShqlT4atfhY02giuvhAMOyDqizqWzPAfJzD4HHOLuJ0TTXwT2cPev583zc2B9d/+OmQ0H7ovmmR/lDwJuTHqs0nFKRCR7jfYcpFOBWe6+AzArmi7mUnTZgojU2TvvwIQJMGYMDB0KCxaoc9TFFRvPsPDXw3OAzcxsIfANYAFQ0VOOzWycmc0zs3mrVq3qUKAiIlJ7WXWQDgcui/6/DDii2EzuPhsoPZyUiEjKnnkG9tkHJk2C738f7rgDBgzIOiqpsReAbfKmtwaW58/g7m+4+1h334VwD1JfYFklhbj7ZHcf5u7D+vbtW2XIIiJSK1l1kPq7+wqA6LVftSvUL3PS8JZNgesHwdQe4XXZlLb5cybAlT1hqoXXORPqH0Ncfi1VE1u5vAra9cYbwxmjJ5+E666DX544hV43lVjvzBFhnbm/mSOS1bOjy5WrR6l1ztipbXr0kFtpZy6wg5kNNrPewDHA9PwZzGzTKA/gBGC2u79R5zhFRKQOajZIg5nNBN5fJCuFp0625+6TgckQru2uRRkiHbZsCswZt+4Boi3PhmkIN+DPmQBLJ62b31vXTad143xcDHH5tVRNbFA6b9W9idp1zRr40Y/g7LNhl11g2jTYrkeZMp/6S/uH266cFTom+QM1FCoc2CHpcuX2jzeeKL7OK5vBCx5Mu3pJ6CTlD9QguPsaM/s6cCvQBFzi7ovNbHyUfyHwYeCvZtZKGLzhK7nlzexKYH9gCzN7Afixu19c52qIiEhKshqk4XFgf3dfYWZbAne5+44l5h1EBTe+gm5+lQZ0/aDwBbtQ87ZwxDPhjECxh55aU3hmUD1iiMuvpWpig9J5/30htl3//W8YPRruuisMyHDuubDBBjExFUvPObbMZ+rUYre6JFiu3P5R9mG5HSgrRZ1lkIYs6DglIpK9RhukYTpwfPT/8cANGcUhUh8tz5VPL/UltyNffjsaQ1x+LVUTW7m8mHadPRt23RUeeAAuvRQmT446R0liqqd67B8iIiICZNdBOgc4yMyeBA6KpjGzAWZ2c26m6LKFfwE7mtkLZvaVomsTaXTNA8unW1Px/FLptYghLr+WqomtXF6J9lvrPfnlL8PIdBtvHDpIxx9fZPlKYq2leuwfIiIiAmTUQXL3V9z9QHffIXp9NUpf7u6H5c032t23dPde7r61rumWTmvns6CpuW1aU3NIh/CQzmJKpdcihrj8WqomtnJ5Rdrvtbc25bMXzeOUU+DII2HuXPjoRyuMqd+BxetRKj0uP265cvtHqWVtg+LpfYaUL0tERKSby+oMkkj10hpxLcl6qi1r8Bh4315t096317rBD4Zf0P6La58hlQ/QUC7OuBgGj4HmQW3zmwdVNkBDR0eTSxLb4OPXnTGxpjA9eEz5uAvadf6yoex2+sPccv/OnHsuXH01bHx3iZHeBo+B1rfbrrf17ZA+YibtPz57rBtoodSocqUGYsilX71Z2+Wu3iykD7+gfYfHNgjpI2YCvQpW2AtGtxTfp3IDNJTbHlmOZigiIpKxTAZpqDXd/NoNFI5qBuHX/eGTK/9CH7eeNMoqHIUsZ/sTw5fcuPw06hJXRrER1iCcoSg3wlqS8gtHkyssOy62cusuNqJcLu6NPwhLJ+EOk+8Yxzf/+gf6b/IS1/ziKvb8yg9CZ2j1kvbL9hkCq58E3itS0V7QtCG0vt4+q2lTeN9upeNZeSewtsg6e0DTxqXX2bMZ3lnePm+96AFNpfKOerFIWZRv774fT+e9hQZpKEfHKRGR7JU6TqmDJJ1TWiOuJVlPGmXFjVKXxih21Y6U19ER1pKUHzeaXFxsHR1Rzpp467/rMf6SC7ni3i/yqY/eypSTxrDFxq/H17krKLXdyrX3BlunNpqhOkil6TglIpK9Usepmj0HqUtbNgUemhhGs2oeGO5JqPVzYqSttEYYS7KeJPPMmQBPTQ5fOq0p3BuSf+YnbhSypKOUFZ7lyT+7U4+R8srVs2z5Jb6oJ61/B7f3Yy9uz+fOncaSF4fwk6N+xMQjzqKpx9qS4XQ5pfaXcu3dSKP3iYiIZED3IFUqd6lPy7OAr3t4pK7Rr6+0RhjrtXl8elxZucuV8r/sL53U9p6ONJR7yGiSOKsdCS2unjUd9a3UR1Xpj7Cr/jWK3U+fy0v/6c+tpxzMj448M3SOupNy+0spvUu8J0qli4iIdDHqIFXqoYltr82HMP3QxGzi6a7SGnGt1BVW+elxZT01ufg62qQnKShGsfta8tOrHSnP1isRYpQeV8+ajoJX6ixX+/R33uvNNy77A6PPu4qPbfMwC87alYM+muAeqk6rwssEV84qs4yVPrvWXc66iYhIt6cOUqV0+UljGDwm3DTevC1g4bUDN5Hz7qvx6XFlJbp0rQ7fOuPiHH5BuAk/fyS4/EEg/N0SIUbpcfVMa5tU4dlVA/nEmbM577Zv8J1Df8NdP9yfrd9XYqCCbq3M/vheifdEqXQREZEuRvcgVap5YIkbmDN4eGR3lxviuRpJt2e5sqyp9A3vlcyThrg2GX5B6VHx4toiSR3S2CaViuK6ZeEhfGHSFaxp7cm0k4/iqD1uqOz+qs6q1HYrp9y2LDlIgz7jRESke9AZpEpl+TBNSV8a2zPJQ17j5knyANGOPmQ0qVo+zDYu9rj8Ug837TOE1kFf4/S//ZTDfnULW2/+AvN+Noyjhv89xFXqYam59HLlNm1aPK9p07LxlFWuvHJ5ueG8C603oPR2KxVLvwPLb0t9xomISDenDlKlGuAyok6vHg9mTbqeNLZn3KVruXkKv+SuN2DdPCNmFn+oZ/7zh0bMLP6w0CTPKEqi3MNYc3WIe5htuQekFnY4mjZdF3vcg1dHLi7afiv3WMzBp5/Pz64/nS/vdzH3/2RPdnj/UqBXiGt0S/H1jm7JK7eIETNh1GvF80a9Fj1stch6Ry4uPbT2sR7WW2w/GDGzfN5RLxZvv6NeLP2Q3RJtxoiZ5fdZfcaJiEg3p+cgSX3V68Gsaa4nDXEPYU0Sa7mHmo5cXH2MtXzQLJSv/9WblX5I6qjXisZ2z5MHMmrSdF591Tj/Syfx5f3/0n7ZwaPLx3xlM/h/2+fbBrDR4NLt/e7rpR/OusmHYx9cWzQe6NjDW0s9hLfPkOLxV/Lw4SrpOUil6TglIpI9PShWqhP3nJ+k6vVg1krWU4/nWsU9hDVJrNU+yBXKb8daPmi2nGO9ovZxh9/e/B1OueoXDO7/PNO+cQQ7b/tw8WXL3WuTxYNi4+7j6sjDW0s9hLdcDKPX1GW/VwepNB2nRESypwfFSscVnjnIPf8GKu8kpfVg1kRllbhxPT+98MxE7rlWUN+zTPUYHTFuO9bjQbMdFW2z19/ahLGT/8L18z7LkbtfyyXjvswmzW+UXi7LmCspt1w8sQ9vrfBHLm9tnP1eRESkAekeJImX6Dk/CSV5mGhaDxxN8mDURnmuVU0fshqJ2471iKEKC5/ZmWGnz+PGBZ/ht1/4NtNO/lz5zlEj6ujDetPcNtbUOPu9iIhIA1IHSeKl+St8khGy0hpFK0nc9XquVdwobUnq3NGR03Li2qOWo5fF1b/MiHHucPFdX2bPM+7n7XfX566J+/PtQ3+P1frquFqMVNfRkQDT3DbbjdPz3ERERMpQB0nidfRX72KSjJCV1ihazdvGp9frrMmIme2/OOcGKIBkdR65uPgockkHaIjbjnExJGnPUuLqP+q1oqO0tYx8jbFj4YQ/Xcy+O/6TB88aysd3vK9t7OXqFRdzufyOtne5upYbPa5cPcptm3LLlSqrwc8WioiIZEkdpFpLa7jqLFXz/JtiBo8JN/0fuza8Fuv4JJknTtKzVT16t52nR+/2v8zHbcc5E8IgBlMtvM6Z0D6eETPDgAO5v44Mz50bRjr3V/hlvVwcSbbjqnvDTf94eF1177q8nc+i6NDWubaKO0v03xVt0wunB49u82X+iaaT2XNP+Otf4UdfuYl/nHII/TZZ1T72uGf6UHiqydbFXOoMTC59gy3bpuem4+q63di2nZntxhafvzDecuml3hPllht+QRiQ4VgPr7l7BvWsIxERkZI0SEMtdZUboft+HJZeBKzNS+wR0htZro3jRuoqHMmxcDpuO6YxiEUa+0pcHLlYSo1iF7f8qntpuw8QplfdGx9jsSHKVy8J6SMXtyt72v1H8OXJ36H3em9y880bcUjP38HKwrKBN56A/zxavMznbwj57QYxcHjqLyHmRT8vvuyin4d5CofrXjkrGs58dvHlVs4uvy0Lh+Qutq+kMVpknKTvDRERkW5Iw3zXUlrDVWetq9SjmDSGHY8b/jqtOOJUG0fc8rUc5jta97trevGDK3/Juf/4Fntu/y+uOXk023z9mdoMxx03vHhHNW9b+ZDclewrhdLY/2pEw3yX1jDHKRGRbkzDfCeR9nNBusqN0GnXox7PHUoqjWHH0xjEIo02ThJH4Zmc/Htq4pav5ZDZ3srzr2zN5/9wDfcv3YuTD/k9vxz9A3r3fK/6dddbR4bkTtKGhQ/izd3X1GhDmYuIiHRyugcpJ3dZTMuzgK+7LKaae4a6yo3QadajFu1cjTSGHU9jEItU2rjU2ZAovdxlbomU+rio/mPk1ocPZdfTFrDohY9wzTeP5vdf/HboHHVkIJCs9dywsvQkCjtHsO5yvzQHURERERF1kP6nFs8F6So3QqdZj0Z7/koaw46nMYhFKm1c6nLZKL2wc5RTKr1Qjw3Kp8cNXFBE69oenHEGHPrLG9ly0xXMO3MYR+8xbd0MuTbssX6JstcntmPYUda7svScNW+VSe9grIWdo/z0tAdRERER6ebUQcqpxeVwaQ1XnbU069Folx2mMex4uaGb04wja2tbyqfHDeVdYNUbW3DoL27hJz+BL36xBw9M+TM7bvVUyCxsw7XvlCi7RHoavMTlfaXS181QJj2mE9sRaex/IiIi8j+6BymneWCJG6urvBxu8JjG+pLbUWnVo1btXI0kdYubJ3+UuFrGUY41lb5ZPw1Jtl3Cocvve2IvPv+Ha3j5zS3405/gK18Bs3OBcztWdqm8coMi5F5L5W+wdcfX29G8jkpj/xMRERFAZ5DW6SqXwzU6tXPtxF1qVfjQ05xcelx+NdsuOrPkDr+/5WT2+9ndrNfrHf71h29xwglgcVfDlSu7XF5cm8Q9Q6kj6+1oXjkduHxRREREOkYdpJzOcIlTV6B2rp24S61GLm7fCcofxS4uv5ptN2Imb2w4kqPP/RvfvuL3fHqXm5h/0XfYdfyFyepWruxyeXFtUi6/o+vtaF5M+1Vy+aJUzswOMbPHzWypmZ1aJH8zM7vOzB42szlm9pGky4rUTJKH0Sd5kHmtY0gyT63FxVBtPlTf1nFlzBwR1p37mzmisvWntY64esaVMWOntvmJB4uqHz0HSURq7uGH4XOfg6efhnPOge9+N8FZI+n0OstzkMysCXgCOAh4AZgLjHb3JXnz/Ap4091/YmYfAs539wOTLFuMjlNStcKHUkM4y53/w1XhA8Bz0rpPMUkMSeaptbgYqs2H6ts6roxio5lCZT+WpbGOuHrGlVFsRF1o+4NsHZU6TukMkojU1KWXwh57wJtvwp13wve+p86RNJzhwFJ3f9rd3wWuAg4vmGcIMAvA3R8DBplZ/4TLiqQvyaiwT00uvmyp9FrE0Aij18bFUG0+VN/WcWWUG800qTTWEVfPuDKqHVG3TtRBEpGa+O9/4YQTYOxY2GsvWLAA9t0366hEitoKeD5v+oUoLd9DwJEAZjYc2BbYOuGyIulLMipsrR8kncbD1ushLoZq86H6tm6EdkqimzycXB0kEUnd0qWhU3TxxTBxItx+O/Tvn3VUIiUVO6dZeP35OcBmZrYQ+AawAFiTcNlQiNk4M5tnZvNWrVpVRbgiJHvAeK0fJJ3Gw9brIS6GavOh+rZuhHZKops8nFwdJBFJ1XXXwW67wXPPwU03wc9+Bk1d63NTup4XgG3yprcGlufP4O5vuPtYd98FOA7oCyxLsmzeOia7+zB3H9a3b98Uw5duKcnIorV+kHQaD1uvh7gYqs2H6ts6row0RjNNYx1x9YwrI27E3AahDpKIpOK998LgC0ceCTvuGC6pO+ywrKMSSWQusIOZDTaz3sAxwPT8Gcxs0ygP4ARgtru/kWRZkZpIMrJorR8kncbD1ushLoZq86H6to4rI43RTNNYR1w948qIGzG3QWgUOxGp2osvwqhRcO+9cNJJ8JvfwHrrZR2VZK2zjGIHYGaHAb8HmoBL3P0sMxsP4O4XmtlewF+BVmAJ8BV3f63UsnHl6TglIpK9UsepnlkEIyJdx8yZcOyx0NICV14JxxyTdUQilXP3m4GbC9IuzPv/X8AOSZcVEZHOS5fYiUiHrF0LZ54Jn/oU9O0Lc+eqcyQiIiKdn84giUjFXn4ZvvhF+Mc/YMwYuOgi2HDDrKMSERERqZ46SCJSkQcegKOPhpdegkmT4Gtf04NfRUREpOvQJXYikog7/PGP4WGvTU1w330wfrw6RyIiItK1qIMkIrFWrw73F33zm3DwwfDgg+FZRyIiIiJdjTpIIlLWokWw++4wbRqccw7ccANstlnWUYmIiIjUhu5BEpGSLr883GO08cYwaxbsv3/WEYmIiIjUViZnkMxsczO73cyejF7b/R5tZtuY2Z1m9qiZLTazk7OIVaQ7evvt0DE67jgYPhwWLFDnSERERLqHrC6xOxWY5e47ALOi6UJrgO+6+4eBPYGTzGxIHWMU6Zaefhr23hsmT4ZTTw0Pgt1yy6yjEhEREamPrDpIhwOXRf9fBhxROIO7r3D3B6P/VwOPAlvVK0CR7mj6dBg6FJYtC/+ffTb01IW4IiIi0o1k1UHq7+4rIHSEgH7lZjazQcCuwANl5hlnZvPMbN6qVavSjFWky1uzBk45BQ4/HLbfPoxSN3Jk1lGJiIiI1F/Nfhs2s5nA+4tkTaxwPRsB1wLfcvc3Ss3n7pOByQDDhg3zSsoQ6c5WrIBRo+Cf/wzPNfrd72D99bOOSkRERCQbNesgufuIUnlm9pKZbenuK8xsS2Blifl6ETpHU9z97zUKVaTbuvNOGD06POfoiitgzJisIxIRERHJVtlL7KLR5kr+VVHudOD46P/jgRuKlG3AxcCj7v7bKsoSkQJr18LPfw4jRoRnGs2dq86RiIiICMSfQZoPOGDAQOC16P9NgeeAwR0s9xzgGjP7SrSeowHMbADwZ3c/DPg48EXgETNbGC13mrvf3MEyRQR49dUwfPdNN8Exx8Cf/gQbbZR1VCIiIiKNoWwHyd0HA5jZhcD0XOfEzA4FSl5CF8fdXwEOLJK+HDgs+v8eQmdMRFIydy4cfTQsXw7nnQcTJoDpXSYiIiLyP0lHsds9/8yNu98C7FebkEQkbe5wwQWwzz7h/3vugZNOUudIREREpFDSDtLLZvZDMxtkZtua2UTglVoGJiLpePPNcH/RSSeFe44efBCGD886KhEREZHGlLSDNBroC1wX/fWN0kSkgS1ZEjpDV18NZ50FM2bA+96XdVQiIiIijSvRMN/u/ipwsplt5O5v1jgmEUnB1Knw1a+GARhuvx0OOCDriEREREQaX6IzSGa2t5ktAZZE0zub2QU1jUxEOuSdd8LgC2PGwNChsGCBOkciIiIiSSW9xO53wMFE9x25+0PAJ2oVlIh0zDPPhIEYJk2C738f7rgDBgzIOioRERGRziPRJXYA7v68tR3yqjX9cESko268MTzfaO1auO46OOKIrCMSERER6XySnkF63sz2BtzMepvZ94BHaxiXiCS0Zg2cdhqMHAmDBsH8+eociYiIiHRU0g7SeOAkYCvgBWAXYEKNYhKRhP79bzjoIDj77DAgw333wXbbZR2ViIiISOeV9BK7Hd19TH6CmX0cuDf9kEQkidmzYdQo+M9/4NJL4fjjs45IREREpPNLegbpjwnTRKTG1q6FX/4yjEy38cbwwAPqHImIiIikpewZJDPbC9gb6Gtm38nL2hhoqmVgItLea6/Bl74E06fD0UfDn/8cOkkiIiIiko64M0i9gY0IHak+eX9vAJ+rbWgikm/+fNhtN7jlFjj3XLj6anWORNJiZoeY2eNmttTMTi2Sv4mZzTCzh8xssZmNzcs72cwWRenfqmvgIiKSurJnkNz9buBuM7vU3Z+tU0wikscdJk+Gb34T+vcP9x7tuWfWUYl0HWbWBJwPHEQYiGiumU139yV5s50ELHH3kWbWF3jczKYAHwS+CgwH3gX+YWY3ufuT9a2FpG7ZFHhoIrQ8B80DYeezYPCY+OUqWb7aMuZMgKcmg7eCNcF242D4BZXPU009kqx/5ghYOWvddL8DYcTM+sYQN09cGUnqEFfGjJ1gdd7HSp8hMHLxuumpvYH38lbYC459d93k1ZtB6+vrpps2hVGvtY0hLs6pTcDavAV6wLF5T+5JUsa1W8E7y9dNrzcAjnoxeX6cJNuz2vdOjKT3IP3ZzDbNTZjZZmZ2a2pRiEhRb70Vnm00fjx88pPw4IPqHInUwHBgqbs/7e7vAlcBhxfM40AfCw8E3Ah4FVgDfBi4391b3H0NcDfw2fqFLjWxbArMGQctzwIeXueMC+lpLV9tGXMmwNJJ4UskhNelk0J6JfNUU48k6y/8wg5heuaI+sUQN09cGUnqEFdGYecIwvSMncL/7TpHhOmpvcO/hR0XCNNXb7ZuOi7Odp0jwvTUpuRlFHZ+IExfu1Wy/DhJtme1750EknaQtnD313MT7v4a0C+1KESknccegz32gClT4Kc/hZtvhi22yDoqkS5pK+D5vOkXorR85xE6Q8uBR4CT3X0tsAj4hJm9z8yagcOAbWofstTUQxOhtaVtWmtLSE9r+WrLeGpyfHqSecqJizHJ+gu/sMel1yKGuHniykhSh7gyCjtHOf9LL+wc0Ta9sOPyvzjz0mPjLOwc0TY9SRmFnZ/C9Lj8OEm2Z7XvnQSSDvO91swGuvtzAGa2LeHXNBGpgauuCs81Wn99uPXW8KwjEakZK5JWeIw7GFgIHABsB9xuZv9090fN7BfA7cCbwEOEM0vtCzEbB4wDGDhwYDqRS220PFdZekeWr7YMb41PTzJPOXExVrv+esUQN0+12yJpHBIvSTumsb1iJD2DNBG4x8wuN7PLgdnA/0stChEB4J134BvfgNGj4WMfgwUL1DkSqYMXaHvWZ2vCmaJ8Y4G/e7AUWAZ8CMDdL3b3oe7+CcKld0XvP3L3ye4+zN2H9e3bN/VKSIqaS3RgS6V3ZPlqy7ASgwnnpyeZp5y4GKtdf71iiJun2m2RNA6Jl6Qd09heMRJ1kNz9H8BQ4GrgGmA3d9c9SCIpevZZ+MQn4Lzz4Dvfgbvugq23zjoqkW5hLrCDmQ02s97AMcD0gnmeAw4EMLP+wI7A09F0v+h1IHAkcGWd4pZa2fksaGpum9bUHNLTWr7aMrYbF5+eZJ5y4mJMsv5+Bxafp1R6LWKImyeujCR1iCujz5Di+f9L71U8P5fetGnx7Pz02DhLfe3vkbyM9QYUnyeXHpcfJ8n2rPa9k0DZDpKZfSh6HQoMJPyi9iIwMEoTkRTccgsMHRruO7r2WvjNb6BXqc9KEUlVNLjC14FbgUeBa9x9sZmNN7Px0WxnAnub2SPALOAUd385yrvWzJYAM4CTovt0pTMbPAaGT4bmbQELr8MnJx8lK8ny1ZYx/ALY/sR1v6xbU5jOH+0ryTzV1CPJ+kfMbP/FvZJR7NKIIW6euDKS1CGujJGL23eS8kexO/Zd2neS8kaxG/Va+w5M4QhzcXEe20r7r/55o9glKeOoF9t3dvJHqYvLj5Nke1b73knA3EvfSmRmf3L3r5rZnUWy3d0PSC2SFA0bNsznzZuXdRgisVpb4Ywz4Gc/C5fUTZsGO+yQdVQi6TCz+e4+LOs4GpGOUyIi2St1nIp7DtJXo9dP1iowke5q5Uo49liYNQu+/OVwad0GG2QdlYiIiEj3VraDZGZHlst397+nG45I93DPPTBqFLz6Klx8ceggiYiIiEj24ob5Hhm99gP2Bu6Ipj8J3AWogyRSAXf47W/hlFNg8ODwbKOdd846KhERERHJibvEbiyAmd0IDHH3FdH0lsD5tQ9PpOt4/XUYOxauvx6OPBIuuQQ22STrqEREREQkX9LnIA3KdY4iLwEfrEE8Il3SwoUwbBjceGM4gzRtmjpHIiIiIo0o7hK7nLvM7FbCsx2c8IyIYiPbiUge93Cm6KSTYIst4O67Ye+9s45KREREREpJ1EFy96+b2WeBT0RJk939utqFJdL5tbTAhAlw2WVw0EEwZQr07Zt1VCIiIiJSTtIzSAAPAqvdfaaZNZtZH3dfXavARDqzJ56Az30OFi2CH/8YTj8dmpqyjkpERERE4iTqIJnZV4FxwObAdsBWwIXAgeWWE+mOpk0Lw3b37g233AIHH5x1RCIiIiKSVNJBGk4CPg68AeDuTxKG/haRyLvvwre+BUcfDTvtBAsWqHMkIiIi0tkk7SC94+7v5ibMrCdhsAYRAZ5/HvbbD849F04+OQzGsM02WUclIiIiIpVKeg/S3WZ2GrCBmR0ETABm1C4skc7j1lthzBh45x245ppwBklEREREOqekZ5BOAVYBjwBfA24GfliroEQ6g9ZWOOMMOPRQ2HJLmDdPnSMRERGRzi72DJKZ9QAedvePAH+qfUgijW/VqnDW6Pbb4bjjYNIkaG7OOioRERERqVZsB8nd15rZQ2Y20N2fq0dQIo3svvvg85+Hl1+GP/0JvvIVMMs6KhERYNkUeGgitDwHzQNh57Ng8Jiso+qe0tgWcetIUsbMEbBy1rrpfgfCiJnJ89OQRj3KmTMBnpoM3grWBNuNg+EXVDbPjJ1g9ZJ1032GwMjFlZURt464tk5SRpxrt4J3lq+bXm8AHPVi8hjS2KfixNUzSTvU+LMu6SV2WwKLzWyWmU3P/aUWhUgn4A6//30YjGG99eBf/4ITTlDnSEQaxLIpMGcctDwLeHidMy6kS32lsS3i1pGkjMIvshCmZ45Ilp+GNOpRzpwJsHRS+DIN4XXppJCedJ7Cjg2E6Rk7JS8jbh1xbZ2kjDiFnSMI09dulSyGNPapOHH1TNIOdfisM/f4wejMbL9i6e5+d2qRpGjYsGE+b968rMOQLuSNN8Kzja69Fg4/HC69FDbdNOuoRBqbmc1392FZx9GIanKcun5Q9IWhQPO2cMQz6ZYl5aWxLeLWkaSMqWV+wTvW4/PTkEY9yrmy57ov0/msCUavSTZPXDskKaPatk5SRpxqY0hjn4oTV88k7ZDiZ12p41TZS+zMbH1gPLA9YYCGi9094VYS6Roefhg+9zl4+mn41a/gu9/VWSMRaUAtJa6CL5UutZPGtohbR2fZ3rWuR7Ev04XpSeaptoxqJSkj7hK+aqW1T5W7/C2unknaoQ77ftwldpcBwwido0OB36RWskgncOmlsMce8OabcOed8L3vqXMkIg2qeWBl6VI7aWyL3puXT4/LbxRxbdGrRLyl0tsp9VW2R4XzlGFNlaXXQtwlfGnouWFl6cXEXv5W6ktUlJ6krevwWRe3Zwxx9y+4+0XA54B9UytZpIH997/h/qKxY2HvvWHBAthXe7+INLKdz4KmguE0m5pDutRXGtui1NVKnjAfws3zxeTS4/LTENcWMd+XY/XYID49bp4+Q4rn59K3G1c8v1R6MdW2dWHnKC69IzGseat4fqn0Yh6aCK0tbdNaW0I6xHfCkrR1HT7r4jpI7+X+SfPSOjPb3MxuN7Mno9fNisyzvpnNiUbQW2xmP0mrfJFyli6FvfaCiy+GH/4QbrsN+vfPOioRkRiDx8DwyeE6fCy8Dp+sUeyykMa2eO/V8ulx+RBGFiv8Upw/4lhcfhri2uLdEvUolV5obUt8etw8Ixe37yTlX742/ALY/sR1ZzGsKUxXMsJcPdq66hiS9LpjxF3+FtcJS9LWdfisixvme2czeyP634ANomkD3N037mC5pwKz3P0cMzs1mj6lYJ53gAPc/U0z6wXcY2a3uPv9HSxTJNZ118GXvgRNTXDTTXDYYVlHJCJSgcFj1CFqFNVui+aBJW5EH5gsPyfuC3g9vqCXa4uk9SglyfJJ5om7l2f4BZUPuV2onp2hjsRgTaUHSEg6Txr7bZK2rvFnXdkzSO7e5O4bR3993L1n3v8d7RwBHE64v4no9YgiZbu7vxlN9or+UhpSRaSt994Lgy8ceSTsuGO4pE6dIxERyUzcZURd5ZLKauuRZPl6tFW1l9DFXeYXlw/hmUfFlEovlOTytrh5ush+m/Q5SGnr7+4rAKLXfsVmMrMmM1sIrARud/cHSq3QzMaZ2Twzm7dq1apaxCxd1Isvwic/Cb/9LZx0Evzzn7DttllHJSIi3VrcZURd5ZLKauuRZPl6tFW1l9DFXeYXlw/hgbCFnaHCB8WWk+Tytrh5ush+m+g5SB1asdlM4P1FsiYCl7n7pnnzvubu7e5DysvfFLgO+Ia7L4orW89BkqRmzoRjj4WWFvjzn+GYY7KOSKTr0HOQStNxSkQkex16DlI13L3kI3XN7CUz29LdV5jZloQzROXW9bqZ3QUcAsR2kETirF0LZ50FP/4xfPjDMG1aeBURERGR7i2rS+ymA8dH/x8P3FA4g5n1jc4cYWYbACOAx+oVoHRdL78Mn/40/OhH4ezRnDnqHImIiIhIkFUH6RzgIDN7EjgomsbMBpjZzdE8WwJ3mtnDwFzCPUg3ZhKtdBkPPABDh8Idd8CFF8Lll8OGFTz/TERERES6tppdYleOu78CtBvWw92XA4dF/z8M7Frn0KSLcofzzgsj1W21Fdx3H+y2W9ZRiYiIiEijyeoMkkjdrF4dBl/45jfh4IPhwQfVORKRtszsEDN73MyWRs/nK8zfxMxm5D28fGxe3rejtEVmdqWZrV/f6EVEJE3qIEmXtmgR7L57GIThnHPghhtgs5LjJYpId2RmTcD5wKHAEGC0mRU+dOQkYIm77wzsD/zGzHqb2VbAN4Fh7v4RoAnQeJgiIp2YOkjSZV1+OQwfDq+/Hu45OuUU6KE9XkTaGw4sdfen3f1d4CrCA83zOdDHzAzYCHgVWBPl9QQ2MLOeQDOwvD5hi4hILejronQ5b78NX/saHHdc6CAtWAD77Zd1VCLSwLYCns+bfiFKy3ce8GFC5+cR4GR3X+vuLwK/Bp4DVgD/cffbihWiB5qLiHQO6iBJl/L007D33jB5Mpx6angQ7JZbZh2ViDQ4K5JW+BT1g4GFwABgF+A8M9vYzDYjnG0aHOVtaGZfKFaIu09292HuPqxv375pxS4iIilTB0m6jOnTwxDey5aF/88+G3pmMk6jiHQyLwDb5E1vTfvL5MYCf/dgKbAM+BDhGX3L3H2Vu78H/B3Yuw4xi4hIjaiDJJ3emjXh/qLDD4fttw+j1I0cmXVUItKJzAV2MLPBZtabMMjC9IJ5niN6PIWZ9Qd2BJ6O0vc0s+bo/qQDgUfrFrmIiKROv69Lp7ZiBYwaBf/8J4wfD7/7HayvAXZFpALuvsbMvg7cShiF7hJ3X2xm46P8C4EzgUvN7BHCJXmnuPvLwMtmNg14kDBowwJgchb1EBGRdKiDJJ3WnXfC6NHhOUdXXAFjxmQdkYh0Vu5+M3BzQdqFef8vBz5VYtkfAz+uaYAiIlI3usROOp21a+HnP4cRI8IzjebOVedIRERERNKhM0jSqbz6ahi++6ab4Jhj4E9/go02yjoqEREREekq1EGSTmPuXDj6aFi+HM4/H048EazY4LwiIiIiIh2kS+yk4bnDBRfAPvuE/++5ByZMUOdIRERERNKnDpI0tDffDPcXnXRSuOfowQdh+PCsoxIRERGRrkodJGlYS5aEztDVV8NZZ8GMGfC+92UdlYiIiIh0ZboHSRrS1Knw1a+GARhuvx0OOCDriERERESkO9AZJGko77wT7i8aMwaGDoUFC9Q5EhEREZH6UQdJGsYzz4SBGCZNgu9/H+64AwYMyDoqEREREelOdImdNIQbbwzPN1q7Fq67Do44IuuIRERERKQ70hkkydSaNXDaaTByJAwaBPPnq3MkIiIiItnRGSTJzL//DaNHw113hQEZ/vAHWH/9rKMSERERke5MHSTJxOzZMGoU/Oc/cNll4fI6EREREZGsqYMkdbV2Lfz61+Gyuu22g9tug49+NOuoRESkS1k2BR6aCC3PQfNA2PksGDym8nmqLaPWZo6AlbPWTfc7EEbMrCzOGTvB6iXrpvsMgZGL040zLoY5E+CpyeCtYE2w3TgYfkHyeiZph6s3g9bX1003bQqjXkseY9zySepxZTP4f9dN2wYwumXddNy2uHYreGf5uun1BsBRL7aNIa4t4mKMy0/SVmmUkTHdgyR189pr8NnPwimnwFFHwdy56hyJiEjKlk2BOeOg5VnAw+uccSG9knmqLaPWCr8IQ5ieOSJ5nIVfyCFMz9gpvTjjYpgzAZZOCl+WIbwunRTSk9QzSTsUdm4gTF+9WbIY45ZPUo/CzhGE6Subw/9x26KwcwRh+tqt1k3HtUVcjHH5SdoqjTIagDpIUhfz58Nuu8Ett4R7ja66CjbeOOuoRESky3loIrS2tE1rbQnplcxTbRm1VvhFuFh6XJyFX8hzSqV3RFwMT00uvlwuPa6eSdqhsHNTmB4XY9zy+fEWyqUXdo5yculx26Kwc5STnx7XFnExxuVD9dszSRkNQB0kqSl3uOgi2HvvMGLd7NnwjW+AWdaRiYhIl9TyXHx6knmqLaMRNEKccTHkziQUKpVeC2m0UyPUI05cjEnqUO327AzthDpIUkNvvRUGXxg/Hj75SXjwQdhzz6yjEhGRLq15YHx6knmqLaMRNEKccTFYU/H8Uum1kEY7NUI94sTFmKQO1W7PztBOqIMkNfLYY7DHHjBlCvz0p3DzzbDFFllHJSIiXd7OZ0FTc9u0puaQXsk81ZZRa/0OjE+Pi7PPkOLrKJXeEXExbDeu+HK59Lh6JmmHpk2Lz5NLj4sxbvn8eAvl0m2D4vm59Lhtsd6A4vn56XFtERdjXD5Uvz2TlNEA1EGS1F11Fey+O7z0Etx6K5x+OvTQniYiIvUweAwMnwzN2wIWXodPbjvKVpJ5qi2j1kbMbP+FuHDEsrg4Ry5u/8U87VHs4mIYfgFsf2LbMwzbn7huVLO4eiZph1Gvte/k5I9CFxdj3PJJ6jG6pX0nKX8Uu7htcdSL7TtJhaPYxbVFXIxx+UnaKo0yGoC5e9YxpG7YsGE+b968rMPodt55B773PTjvvHDP0dVXw9ZbZx2ViGTFzOa7+7Cs42hEOk6JiGSv1HFKv+tLKp59Fj7xidA5+s534K671DkSERERkc5HD4qVqt1yC3zhC2GUumuvhSOPzDoiEREREZGO0Rkk6bDW1nB/0WGHhbNF8+apcyQiIiIinZvOIEmHrFwJxx4Ls2bBl78cLq3boMQALSIiIiIinYU6SFKxe+6BUaPg1Vfhkktg7NisIxIRERERSYcusZPE3OE3v4H994fmZrj/fnWORERERKRr0RkkSeT110Nn6Prr4aij4OKLYZNNso5KRERERCRdOoMksRYuhGHD4MYb4Xe/g7/9TZ0jEelazOwQM3vczJaa2alF8jcxsxlm9pCZLTazsVH6jma2MO/vDTP7Vt0rICIiqdEZJCnJPdxjdNJJsMUWcPfd4QGwIiJdiZk1AecDBwEvAHPNbLq7L8mb7SRgibuPNLO+wONmNsXdHwd2yVvPi8B1da2AiIikSmeQpKiWlnBJ3QknhAfALligzpGIdFnDgaXu/rS7vwtcBRxeMI8DfczMgI2AV4E1BfMcCDzl7s/WOmAREakddZCknSeegD33hL/+FX784/Ag2L59s45KRKRmtgKez5t+IUrLdx7wYWA58AhwsruvLZjnGODKUoWY2Tgzm2dm81atWlV91F3ZnAlwZU+YauF1zoSsI2pv2RS4fhBM7RFel02pLL+rSGNbxbXVzBFh/bm/mSPSiLytuHokqWet94lGaId6rKMB3juZdJDMbHMzu93MnoxeNyszb5OZLTCzG+sZY3c1bVq432j58tAxOuMMaGrKOioRkZqyImleMH0wsBAYQLik7jwz2/h/KzDrDfwf8LdShbj7ZHcf5u7D+upXp9LmTIClk8Bbw7S3hulG6iQtmwJzxkHLs4CH1znj1n2Ri8vvKtLYVnFtNXMErJzVdpmVs9LtHMTVI0k9a71PNEI71GMdDfLeyeoM0qnALHffAZgVTZdyMvBoXaLqxt59F771LTj6aNhpp3BJ3cEHZx2ViEhdvABskze9NeFMUb6xwN89WAosAz6Ul38o8KC7v1TTSLuDpyZXlp6FhyZCa0vbtNaWkJ4kv6tIY1vFtVVhpyCnVHpHxNUjST1rvU80QjvUYx0N8t7JqoN0OHBZ9P9lwBHFZjKzrYFPA3+uT1jd0/PPw377wbnnwsknh8EYttkmfjkRkS5iLrCDmQ2OzgQdA0wvmOc5wj1GmFl/YEfg6bz80ZS5vE4qkPvlOWl6FlqeK58el99VpLGtGqGt4uqRpJ5dYZ9IY3tWu44GaaesOkj93X0FQPTar8R8vwd+ABRe592Oru3umFtvhV13hUWL4Jpr4Pe/h969s45KRKR+3H0N8HXgVsIVC9e4+2IzG29m46PZzgT2NrNHCFc+nOLuLwOYWTNhBLy/1z/6LshKXNddKj0LzQPLp8fldxVpbKtGaKu4eiSpZ1fYJ9LYntWuo0HaqWYdJDObaWaLivwVjgxUavnPACvdfX6S+XVtd2VaW8P9RYceCltuCfPmhcvrRES6I3e/2d0/6O7buftZUdqF7n5h9P9yd/+Uu3/U3T/i7lfkLdvi7u9z9/9kFX+Xst24ytKzsPNZ0NTcNq2pOaQnye8q0thWcW3V78Diy5VK74i4eiSpZ633iUZoh3qso0HeOzXrILn7iOggUvh3A/CSmW0JEL2uLLKKjwP/Z2bPEIZcPcDMrigyn1Ro1arQMfrJT+CLX4QHHoAdd8w6KhEREWD4BbD9iW1/vd/+xJDeKAaPgeGToXlbwMLr8MkhPUl+V5HGtoprqxEz23cC+h0Y0tMSV48k9az1PtEI7VCPdTTIe8fcCwfqqUOhZr8CXnH3c6Inlm/u7j8oM//+wPfc/TNJ1j9s2DCfN29eKrF2NffdB5//PLz8Mpx3HnzlK2DFxm8SEamSmc1392FZx9GIdJwSEcleqeNUVvcgnQMcZGZPEq7bPgfAzAaY2c0ZxdSluYf7i/bbD9ZbD/71r/AQWHWORERERETW6ZlFoe7+CtFoQAXpy4HDiqTfBdxV88C6qDfegC9/Ga69Fg4/HC69FDbdNOuoREREREQaT1ZnkKROHn44PPj1+uvhV7+C665T50hEREREpJRMziBJfVx6KZx4Imy2Gdx5J+y7b9YRiYiIiIg0Np1B6oL++99wf9HYsbD33rBggTpHIiIiIiJJqIPUxSxdCnvtBRdfDD/8Idx2G/Tvn3VUIiIiIiKdgy6x60Kuuw6+9CXo2RNuugkOazfchYiIiIiIlKMzSF3Ae+/Bd78LRx4ZHvj64IPqHImIiIiIdITOIHVyL74Io0bBvffC178Ov/51eM6RiIiIiIhUTh2kTmzmTDj2WGhpgSuvhGOOyToiEREREZHOTZfYdUJr18KZZ8KnPgV9+8LcueociYiIiIikQWeQOpmXX4YvfhH+8Q8YMwYuugg23DDrqEREREREugZ1kDqRBx6Ao4+Gl16CCy+EcePALOuoRERERES6Dl1i1wm4wx//GB722tQE990HX/uaOkciIiIiImlTB6nBrV4d7i/65jfh4IPDEN677ZZ1VCIiIiIiXZM6SA1s0SLYfXeYNg3OOQduuAE22yzrqEREREREui7dg9SgLr88XEa38cZwxx2w335ZRyQiIiIi0vXpDFKDefvt0DE67jgYPhwWLFDnSERERESkXtRBaiBPPw177w2TJ8Opp4YHwW65ZdZRiYiIiIh0H7rErkFMnx7OGpnBjBnwmc9kHZGIiIiISPejDlLG1qyBiRPhl78Mo9P97W8weHDWUYmISJe0bAo8NBFanoPmgbDzWTB4TNZRtddZ4iwnSR3i5pkzAZ6aDN4K1gTbjYPhF6QfRzlpxBC3jrj8mSNg5ax10/0OhBEz103P2AlWL1k33WcIjFxcWQxxZVy7FbyzfN30egPgqBfbllFtnHHLT+0NvJdXYC849t22MdRjn4pTjzJqTJfYZWjFCjjggNA5OvFEuOcedY5ERKRGlk2BOeOg5VnAw+uccSG9kXSWOMtJUoe4eeZMgKWTwpdMCK9LJ4X0NOMoJ40Y4tYRl1/YaYAwPXNE+L+w0wFhesZOyWOIK6OwcwRh+tqt1k1XG2fc8u06R4Tpqb3XTdZjn4pTjzLqQB2kjNx5J+y6K8yfD1dcARdcAOuvn3VUIiLSZT00EVpb2qa1toT0RtJZ4iwnSR3i5nlqcvF1l0rvaBzlpBFD3Dri8gs7DTm59MJOR05+erVlFHaOcvLTq40zbvl2nSPap9djn4pTjzLqQB2kOlu7Fn7+cxgxIjzTaO5cGNPJrhoQEZFOqOW5ytKz0lniLCdJHeLmyf0CX6hUekfjKCeNGOLWkUYZ1cbQVdRjn4rTRdpaHaQ6evVV+L//C/ccff7zoXM0ZEjWUYmISLfQPLCy9Kx0ljjLSVKHuHmsqXh+qfSOxlFOGjHErSONMqqNoauoxz4Vp4u0tTpIdTJ3LgwdCrfdBuefD1OnwkYbZR2ViIh0GzufBU3NbdOamkN6I+kscZaTpA5x82w3rvi6S6V3NI5y0oghbh1x+f0OLJ6fS+9T4pfm/PRqy1hvQPH8/PRq44xbnl7F8/PT67FPxalHGXWgDlKNuYf7i/bZJ/x/zz0wYUIYzltERBqDmR1iZo+b2VIzO7VI/iZmNsPMHjKzxWY2Ni9vUzObZmaPmdmjZrZXfaNPaPAYGD4ZmrcFLLwOn9x4o8N1ljjLSVKHuHmGXwDbn9j2TMv2J1Y2Gli1bZlGDHHriMsfMbN95yF/dLeRi9t3PgpHh6u2jKNebN9JKhzFrto445Y/9l3ad5IKRrGrxz4Vpx5l1IG5e9YxpG7YsGE+b968rMPgzTdh3Di48ko47DD461/hfe/LOioRkfows/nuPizrOOKYWRPwBHAQ8AIwFxjt7kvy5jkN2MTdTzGzvsDjwPvd/V0zuwz4p7v/2cx6A83u/nq5MhvlOCUi0p2VOk7pDFKNLFkCw4fD1VfDWWeFh7+qcyQi0pCGA0vd/Wl3fxe4Cji8YB4H+piZARsBrwJrzGxj4BPAxQDu/m5c50hERBqbOkg1MHUq7L47vPIK3H47nHYa9FBLi4g0qq2A5/OmX4jS8p0HfBhYDjwCnOzua4EPAKuAv5jZAjP7s5ltWKwQMxtnZvPMbN6qVatSr4SIiKRDX9tT9M474f6iMWNgt91gwYLwIFgREWloxe4KLbz+/GBgITAA2AU4Lzp71BMYCkxy912Bt4B29zABuPtkdx/m7sP69u2bUugiIpI2dZBS8swzYSCGSZPg+9+HO+6AASUGPRERkYbyArBN3vTWhDNF+cYCf/dgKbAM+FC07Avu/kA03zRCh0lERDopdZBScOONYQjvJ5+E66+HX/4SevbMOioREUloLrCDmQ2OBlk4BpheMM9zwIEAZtYf2BF42t3/DTxvZjtG8x0ILEFERDotfY2vwpo18KMfwdlnw667wrRp8IEPZB2ViIhUwt3XmNnXgVuBJuASd19sZuOj/AuBM4FLzewRwiV5p7j7y9EqvgFMiTpXTxPONomISCelDlIH/fvfMHo03HUXfPWr8Ic/wPrrZx2ViIh0hLvfDNxckHZh3v/LgU+VWHYh0PDDmYuISDLqIHXA7NkwahT85z9w2WVw3HFZRyQiIiIiImnQPUgVWLs23F90wAGw8cbwwAPqHImIiIiIdCU6g5TQa6/Bl74E06fD5z8Pf/pT6CSJiIiIiEjXoQ5SAvPnw9FHwwsvhHuNvv51sGJPzRARERERkU5Nl9iV4Q4XXQR77x1GrJs9G77xDXWORERERES6KnWQSnjrrXB/0fjx8MlPwoMPwp57Zh2ViIh0a8umwPWDYGqP8LpsStYRNSa1U3rmTIAre8JUC69zJrTNn7FTyMv9zdip/Tri5ql2e80c0Xb9M0dUtnySGJLEGNdWcflxZcQtn0YMSaTR3g1OHaQiHnsM9tgDpkyBn/4Ubr4Zttgi66hERKRbWzYF5oyDlmcBD69zxunLfyG1U3rmTIClk8Bbw7S3huncl+oZO8Hqgucir17StgMUN0+122vmCFg5q23aylmVfWmPiyFJjHFtFZcfV0bc8mnEkEQa7d0JZNJBMrPNzex2M3syet2sxHzPmNkjZrbQzObVI7arroLdd4eXXoLbboPTT4ce6kaKiEjWHpoIrS1t01pbQrqso3ZKz1OTy6cXdnxy8tPj5ql2exV+WY9LLyYuhiQxxrVVXH5cGXHLpxFDEmm0dyeQ1Vf/U4FZ7r4DMCuaLuWT7r6Lu9f8IXzPPhsuq/vYx2DBAhjRtTrDIiLSmbU8V1l6d6V2Sk/uTEPS9I5ohO0VF0OSGOPaKi4/rowk26LaGOR/suogHQ5cFv1/GXBERnG0se22MGsW3HUXbL111tGIiIjkaR5YWXp3pXZKjzVVlt4RjbC94mJIEmNcW8Xlx5WRZFtUG4P8T1YdpP7uvgIgeu1XYj4HbjOz+WY2rh6B7bsv9OpVj5JEREQqsPNZ0NTcNq2pOaTLOmqn9GxX4qtXLr3PkOL5+elx81S7vfodWFl6MXExJIkxrq3i8uPKiFs+jRiSSKO9O4GadZDMbKaZLSryd3gFq/m4uw8FDgVOMrNPlClvnJnNM7N5q1atqjp+ERGRhjJ4DAyfDM3bAhZeh08O6bKO2ik9wy+A7U9sewZi+xNDOsDIxe07QH2GhPScuHmq3V4jZrb/ct7vwJCeVFwMSWKMa6u4/Lgy4pZPI4Yk0mjvTsDcvf6Fmj0O7O/uK8xsS+Aud98xZpkzgDfd/ddx6x82bJjPm1eXMR1ERKQEM5tfj/tHOyMdp0REslfqOJXVJXbTgeOj/48Hbiicwcw2NLM+uf+BTwGL6hahiIiIiIh0O1l1kM4BDjKzJ4GDomnMbICZ3RzN0x+4x8weAuYAN7n7PzKJVkREREREuoWeWRTq7q8A7e7mcvflwGHR/08DO9c5NBERERER6cb0CFQREREREZGIOkgiIiIiIiIRdZBEREREREQi6iCJiIiIiIhE1EESERERERGJZPKg2Fozs1XAsxUssgXwco3C6SzUBoHaQW0AaoOcatthW3fvm1YwXUmC41Rn2AcVY3o6Q5yKMT2dIc7OECPU6DjVJTtIlTKzed39ae9qg0DtoDYAtUGO2iE7naHtFWN6OkOcijE9nSHOzhAj1C5OXWInIiIiIiISUQdJREREREQkog5SMDnrABqA2iBQO6gNQG2Qo3bITmdoe8WYns4Qp2JMT2eIszPECDWKU/cgiYiIiIiIRHQGSUREREREJNItO0hmtrmZ3W5mT0avm5WY7xkze8TMFprZvHrHWUtJ2yCat8nMFpjZjfWMsR6StIOZrW9mc8zsITNbbGY/ySLWWknYBtuY2Z1m9mjUBidnEWutVPCZcImZrTSzRfWOsVbM7BAze9zMlprZqUXyzcz+EOU/bGZDs4izq4hr72ie/aPjzmIzu7uSZRsgxrodNxPsu9+P4lhoZovMrNXMNk+ybIPE2EhtuYmZzcg7Do5NumyDxFiXtkwQ42Zmdl30WTrHzD6SdNkGirNebVn2eFvu2JRKW7p7t/sDfgmcGv1/KvCLEvM9A2yRdbxZtkGU/x1gKnBj1nFn0Q6AARtF//cCHgD2zDr2OrfBlsDQ6P8+wBPAkKxjr2cbRHmfAIYCi7KOOaV6NwFPAR8AegMPFW5X4DDgluh9sCfwQNZxd9a/hO29KbAEGBhN90u6bNYxRv/X5bhZaXsAI4E7Gq0tS8XYaG0JnJb7XAT6Aq9G8zZMW5aKsV5tmTDGXwE/jv7/EDCrnvtktXHWeb8se7ylxLEprbbslmeQgMOBy6L/LwOOyC6UzCRqAzPbGvg08Of6hFV3se3gwZvRZK/oryvdvJekDVa4+4PR/6uBR4Gt6hVgHSR6P7j7bMJBt6sYDix196fd/V3gKkJb5Dsc+Gv0Prgf2NTMtqx3oF1EkvY+Fvi7uz8H4O4rK1g26xjrqdL2GA1c2cFls4ixnpLE6UAfMzNgI8Ln4JqEy2YdY70kiXEIMAvA3R8DBplZ/4TLNkKcdZPgeFvq2JRKW3bXDlJ/d18B4Ysf0K/EfA7cZmbzzWxc3aKrj6Rt8HvgB8DaOsVVb4nawcJlhguBlcDt7v5A/UKsuaT7AgBmNgjYlXAmrauoqA26kK2A5/OmX6B9xzfJPJJMkrb8ILCZmd0VHXuOq2DZrGOE+h03E7eHmTUDhwDXVrpshjFCY7XlecCHgeXAI8DJ7r424bJZxwj1acskMT4EHAlgZsOBbYGtEy7bCHFC43w3LlWPVNqyZ1WhNTAzmwm8v0jWxApW83F3X25m/YDbzeyxqEfbKVTbBmb2GWClu883s/1TDK2u0tgX3L0V2MXMNgWuM7OPuHunuQ8lpfcDZrYR4QD+LXd/I43Y6iWtNuhirEha4dnRJPNIMknasiewG3AgsAHwLzO7P+GyaehwjO7+BPU7blbSHiOBe90992t0I7VlTmGM0FhteTCwEDgA2C6K558Jl01Dh2OMjlX1aMskMZ4DnBv94PoIsIBwlquen7PVxAmN8924VD1Sacsu20Fy9xGl8szsJTPb0t1XRKfjil4e4O7Lo9eVZnYd4bRdp+kgpdAGHwf+z8wOA9YHNjazK9z9CzUKuSbS2Bfy1vW6md1F+KWv03SQ0mgDM+tF6BxNcfe/1yjUmklzP+hCXgC2yZvemvDra6XzSDJJ2/tld38LeMvMZgM7J1w26xifqONxs5L2OIa2l641UlvmFMZYz+8gSeIcC5zj4SaPpWa2jHBvSiO1ZakY59SpLWNjjDprYyEMMgAsi/6a45ZtkDgb6btxqXr0LpFeke56id104Pjo/+OBGwpnMLMNzaxP7n/gU3SiL8QJxLaBu/8/d9/a3QcRPrzv6GydowSS7At9ozNHmNkGwAjgsXoFWAdJ2sCAi4FH3f23dYytXmLboIuaC+xgZoPNrDfhfT69YJ7pwHHRiEF7Av/JXY4oFUvS3jcA+5pZz+iyqz0I9/wlWTbTGOt83EzUHma2CbAfbd/TjdSWRWNswLZ8jnDGEAv3ouwIPJ1w2UxjrGNbxsZoZptGeQAnALOjzki92rGqOBvsu3GpY1M6bek1HoWiEf+A9xFuPnsyet08Sh8A3Bz9/wHCNZgPAYuBiVnHXe82KJh/f7rmKHZJ9oWPEU4vP0z4IPhR1nFn0Ab7EE5RP0y4hGEhcFjWsdezDaLpK4EVwHuEX6++knXsKdT9MMKohE/lPueA8cD46H8Dzo/yHwGGZR1zZ/6La+9o+vuEUeIWES5nLblsI8VY7+Nmwji/BFyVZNlGirHR2jL6LLwt+gxYBHyh0dqyVIz1bMsEMe5FOM48Bvwd2Kze7VhNnHVuy3bHWxIem9JoS4tWJCIiIiIi0u1110vsRERERERE2lEHSUREREREJKIOkoiIiIiISEQdJBERERERkYg6SCIiIiIiIhF1kCRTZvY+M1sY/f3bzF7Mm+4dv4bU49nfzG6scRlv1nL9tWJmu0QPDU5jXTVvZxGRLOi41nmkeVyTrqVn1gFI9+burwC7AJjZGcCb7v7rXL6Z9XT3NdlEJwV2AYYBN2cch4hIw9JxrVPZBR3XpAidQZKGY2aXmtlvzexO4BdmdoaZfS8vf5GZDYr+/4KZzYl+mbvIzJqKrG93M7vPzB6K5u1jZuub2V/M7BEzW2Bmnyyy3OZmdr2ZPWxm95vZx6L0M8zsEjO7y8yeNrNvRulnmtnJecuflcsrUU8zs19F9XnEzEZF6Vua2eyoTovMbF8za4raJTfvt4usr7+ZXRfV8yEz2ztK/0603CIz+1aUNsjMFuUt+73oQE5Ur19EbfVEVH5v4KfAqCiuUWa2X96vogtyT9euVJl27mtmt5vZg9G2fdbMtuhIGSIiWdJxrXsd16Tz0xkkaVQfBEa4e2vuA66QmX0YGAV83N3fM7MLgDHAX/Pm6Q1cDYxy97lmtjHwX+BkAHf/qJl9CLjNzD5YUMRPgAXufoSZHRCtd5co70PAJ4E+wONmNgm4mPDE6XPNrAdwDDC8TB2PjNa3M7AFMNfMZgPHAre6+1nRgbE5mm8rd/9IVK9Ni6zvD8Dd7v7ZaLmNzGw3YCywB+Gp0w+Y2d3Aa2XiAujp7sMtXHrwY3cfYWY/Ijyp+utRDDOAk9z9XjPbCHg7Zp2llGrnHwN3uPvZZnYIMK6D6xcRaQQ6rnWf45p0cuogSaP6m7u3xsxzILAb4QMYYANgZcE8OwIr3H0ugLu/AWBm+wB/jNIeM7NnCQevfPsAR0Xz3GHhuvJNoryb3P0d4B0zWwn0d/dnzOwVM9sV6E84CL1SJv59gCujer4UfcDvDswFLjGzXsD17r7QzJ4GPmBmfwRuAm4rsr4DgOOieFuB/0T1vM7d34rq/XdgX2B6mbggHBAB5gODSsxzL/BbM5sC/N3dX4hZZyml2nkf4LNR+j/MLO7gJyLSyHRc6z7HNenkdImdNKq38v5fQ9t9df3o1YDL3H2X6G9Hdz+jYD0GeJH1W4IYis2TW9c7eWmtrPux4c/Alwi/bl3SgfXj7rOBTwAvApeb2XHu/hrhF7m7gJOicpIoVc9SbZqTq19+3QrjPAc4gXAAvz/6xXJdwWYn5V2qMKDCGL1M7CIinZGOa538uCbdhzpI0hk8AwwFMLOhwOAofRbwOTPrF+VtbmbbFiz7GDDAzHaP5uljZj2B2YTLFoguQRgIPF6wbP48+wMv536pK+M64BDCL2a3xsw7m3Dtc5OZ9SUcPOZEdVjp7n8iXN4w1MK9Nz3c/Vrg9Fx7FJgFnBjF2xRddjEbOMLMms1sQ8IZmX8CLwH9ol8P1wM+ExMrwGrCpRdEZWzn7o+4+y+AeYTLM/7H3c/PO8gvj2mHYu18D/D5KP1TwGYJYhQR6QyeQce1Tndck+5Dl9hJZ3AtcJyZLSScpn8CwN2XmNkPCddZ9wDeI/wK9WxuQXd/18JNon80sw0I12mPAC4ALjSzRwi/On3J3d+JLmnIOQP4i5k9DLQAx8cFGpV3J/B6gksprgP2Ah4i/IL3A3f/t5kdD3zfzN4D3iRcXrBVFEvuR43/V2R9JwOTzewrhF/ITnT3f5nZpcCcaJ4/u/sCADP7KfAAsIxwwI1zJ3BqtB3OBvaxcBNwK7AEuCXBOoo5g+Lt/BPgymj73Q2sIBzMREQ6Ox3XuvZxTTo5cy92llZEOiL6oH8QONrdn8w6ns4s+gWw1d3XmNlewCR33yXjsEREuhUd16Q70hkkkZSY2RDgRsLNozqIVG8gcE10cH4X+GrG8YiIdCs6rkl3pTNIIiIiIiIiEQ3SICIiIiIiElEHSUREREREJKIOkoiIiIiISEQdJBERERERkYg6SCIiIiIiIhF1kERERERERCL/HwsUWzmEeqY3AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))\n",
    "\n",
    "ax1.scatter(y_test, lr_subregions_test_pred, c=\"orange\")\n",
    "\n",
    "p1 = max(max(lr_subregions_test_pred), max(y_test))\n",
    "p2 = min(min(lr_subregions_test_pred), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Linear Ridge regression on subregions - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(lr_subregions_test_pred), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title='Linear ridge regression on subregions - natural scale')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eb5252fc",
   "metadata": {},
   "source": [
    "#### LRR: with features - performance - predictor: subregions\n",
    "\n",
    "Linear ridge regression on subregion data has mixed performance, as compared to the baseline. The model was originally optimized for mean absolute error, which performs worse than the naive model. However, the R-squared is considerably better than the naive model, and the RMSE is marginally better. Test and train results for MAE are equal, indicating that the model is an accurate predictor."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "600c5bd9",
   "metadata": {},
   "source": [
    "#### LRR: with features - fit model - predictor: pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "id": "013aebc9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 40 candidates, totalling 200 fits\n",
      "Elapsed time: 0.99 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "lr_all_pests_mod = lr_grid_w_features.fit(X_train_all_pests, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6b7747c1",
   "metadata": {},
   "source": [
    "#### LRR: with features - CV score distribution - predictor: pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "id": "10cb1ae6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.07 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f9438c4c670>"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEICAYAAACuxNj9AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAAAgZElEQVR4nO3de7wdVXn/8c9XAoQkICE5YCDRYEVMwNMYY6IGISUIEZAABcSKBmJKQbFArTVgqdCW3y+lCBGpAi0RUASCgsRqEQgi0iKYQIxATMFCkkOOuUm4Yzn69I9ZicNm7zP7XPbt5Pt+vfbrzJ41l2fWmb2fPWtm1igiMDMz684bGh2AmZk1PycLMzMr5GRhZmaFnCzMzKyQk4WZmRVysjAzs0JOFg0m6RpJ/5iGPyBpZZ3WG5LeVo91lVn32LT+QY1Yf281ss7S+qveP+q8L50v6Zv1WFdfSbpH0pw0fLKk+xodU6twsqiCpKckvSzpBUnrJH1d0rD+Xk9E/CQi9q0iHu/kLaY/vlCr3T96Om095X8cbYvyyarVOFlU78MRMQyYCLwH+NvSCVrtl/JApkzL7N+tFq9te7xz9lBEPA38B7A/bG2a+LSkx4HH07gjJS2TtFnSf0lq3zK/pHdJekjS85JuAgbnyqZJ6si9HyPpFkkbJG2SdLmkccAVwPvSkc7mNO2Oki6WtDod/Vwhaafcsj4nqVPSWkmzK22fpBMlLSkZd7akRWn4cEmPpfiflvTXFZbzBkl/K2mVpPWSrpP0xpLJZqd4OiV9NjfvZElLJD2XtuWSXNl7U51ulvRzSdNyZfdIulDSfwIvAecWbEu/1FkRSTOAc4GPpP/ZzyvE+1ZJp0haker3fyT9RW45pfvHU5L+WtJySc9KuknS4J5Om8r/Jretc9RNk5ukvSX9OMV4JzCypPxmSb9O67lX0n5p/KnAx4C/SfXwvTR+rqRfpeU9JumYPtT1cEn/nj4zz6Th0b1Yzpam0lMr7KNvyMW9SdJCSbulssGSvpnGb5b0M0l7SLoQ+ABwedr+y5W5NH1Gnk3/n/17u/01FRF+FbyAp4BD0vAY4FHgH9L7AO4EdgN2IjvyWA9MAbYDZqX5dwR2AFYBZwPbA8cBrwL/mJY1DehIw9sBPwcuBYaSJZUDUtnJwH0lMc4HFqU4dga+B/z/VDYDWEeW4IYC30pxv63Mtg4Bngf2yY37GXBiGu4EPpCGhwMTK9TZbOAJ4K3AMOAW4BupbGxa/w0pnncCG3J1fD/w8TQ8DHhvGt4L2AQcTvZD54PpfVsqvwdYDewHDALeWLAt/VJnVe5D5wPfLBlXGu/2wBHAHwECDiJLIhNL94/cfvkgsGfahhXAab2Ydgbw6xTHEOAb3W1r+v9cQrZPH5jq+Ju58tmpPndMdbwsV3YNaX/PjTs+xfUG4CPAi8CoXtbzCOBP03bsDNwMfLekzudU+hzlphtL9/voWcBPgdFpO68Ebkhlf5H2pSFkn+N3A7uUrj+9PwxYCuya/ufjervttX41PIBWeKUP2gvAZrIv+68CO6WyAA7OTfs1UiLJjVuZPvgHAmsB5cr+i/LJ4n1p5xxUJp7X7ORpJ3sR+KPcuPcBT6bhBcC8XNnbC74Mvgn8XRreJ30ZDEnvV6cPwy4FdbYY+FTu/b5kiXFQ7oP4jlz5RcDVafhe4AJgZMkyP09KOLlxPwRmpeF7gL+vZlv6u86q2IfOp3yy+PuC+b4LnFm6f+T2y5NK6vCKXky7gJQk0/u3VdpW4M1AFzA0N+5bpduWK9s1LeuN6f01lCSLMvMsA2b2pp7LLGsC8ExJnfckWVTaR1cA03Nlo/jD/j2b7HPdXma5W9ef3h8M/DfwXuAN/bHNtXq5Gap6R0fErhHxloj4VES8nCtbkxt+C/DZdPi5WVkz0RiyX057Ak9H2kuSVRXWNwZYFRFdVcTWRvYFuDS3ztvTeNJ68zFWWucW3wI+mob/jOyX2Uvp/Z+S/bJflZoi3ldhGXuWrGcV2Qdpj9y40pj2TMOfJPty/mU6hD8yjX8LcHxJ3R5A9kEtt8zutqXf6kzZlUcvpNejlaar4DXxSvqQpJ9K+k2K6XBKmnlK/Do3/BLZkVhPpy3d1tI6zNuT7Mv3xdy4rXUjaTtJ81LzzHNkSQq62QZJn9Afmm03kx3NlZ0+V88vSHpzmfIhkq5U1vz5HNkPj10lbdfNNnWn0j76FuDWXMwrgN+R7d/fIPsRc2NqwrpI0vblFh4RdwOXA/8CrJN0laRdehlrTTlZ9I/8l/8a4MKUWLa8hkTEDWRNOHtJUm761+3wueW8WeVPmpd2FbwReBnYL7fON0Z2Qp603jFVrHOLO4CRkiaQfdF+a+uKI34WETOB3cl+9S6ssIy1ZB+o/Dq7yJp2tiiNaW1ax+MR8dG0jn8Cvi1pKFmdfKOkbodGxLzcckrrptK29FudRXbl0bD02q/SZEXjJe0IfAe4GNgjInYFfkB2FFRLnWTNKVuMqTRhmnZ4+n9ska+bPwNmAoeQNQOOTeO3bMNr6kHSW4B/Bc4ARqRtfoQK25yr52ERsbrMJJ8lO4qdEhG7kB3N59ffU2X3UbJ98UMl++LgiHg6Il6NiAsiYjzwfuBI4BNbNqHMNl0WEe8mawZ8O/C5XsZaU04W/e9fgdMkTUknr4ZKOkLSzmRtvV3AX0oaJOlYYHKF5TxI9sGcl5YxWNLUVLYOGC1pB4CI+H1a76WSdgeQtJekw9L0C4GTJY2XNAT4YncbkI5mvg38M1n79p1pmTtI+pikN0bEq8BzZL+myrkBOFvZydBhwP8Dbio5Ujov/RLcDzgFuCmt5yRJbWm7Nqdpf0fWpPRhSYelX7CDlZ3IrXgCs9K29HedVWEdMFbdX/G0A1n79wagS9KHgEP7uN5qLAROkTQubevfVZowIlYBS4AL0v5wAPDh3CQ7A78lO5c0hOz/nreO7DzWFkPJvkA3AEg6hXTxSC/tTPYjYHM64dzX/1vZfZTsIpMLU7JDUpukmWn4TyS9Mx3NPEfWPLXlc/Ka7Zf0nvRdsT1Zs+grVP5MNZSTRT+LiCXAn5MdWj5DdpL35FT2v8Cx6f0zZCfzbqmwnN+RfQjfRnaeoCNND3A32Un2X0vamMZ9Pq3rp+nw+y6yX1hExH+QnWi8O01zdxWb8i2yX4c3l3zBfxx4Kq3jNOCkCvMvIDscvxd4kuxD8JmSaX6c4lkMXBwRd6TxM4BHJb0AfJnshPQrEbGG7FfruWRfLmvIfoUV7ceVtqW/66w7N6e/myQ9VG6CiHge+EuyL+9nyH6lL+rjegulbb0M+BHZtt6fin5bYZY/I7uA4zdkX8bX5cquI2uueRp4jOwkcN7VwPjUfPPdiHgM+FJa5zqyE8n/2YfNmU92ocnGtO7b+7AsqLyPfpnsf3OHpOfTuqaksjeR/UB5jqx56sdkP3S2zHecsiu1LgN2IfvR8gxZvW0iO7JsOnpt87mZbeuUXZ79CLBjlefMBhxJY8l+5Gy/rdZBKR9ZmBmSjknNSsPJzhN9z1+SludkYWaQXQ69AfgVWZv56Y0Nx5qNm6HMzKxQzY4sJC1Qdgv7I7lx/yzpl8puab9V0q65snMkPSFpZe6KFDMzawI1O7KQdCDZXc/XRcSWfpQOBe6OiC5J/wQQEZ+XNJ7sUsvJZDe93AW8PV0RVNHIkSNj7NixNYnfzJrXyk1Z7+v7jmi6jnVbwtKlSzdGRFvxlH9Qs15SI+LedEVBftwdubc/JesbCbLLIW+MiN8CT0p6gixx3E83xo4dy5IlS7qbxMwGoGnXTAPgnpPvaWgcrUpSUS8Or9PIE9yzyXpvhayDuPxt9R1p3Oso6wVyiaQlGzZsqHGIZmYGDUoWkr5Adifz9VtGlZmsbPtYRFwVEZMiYlJbW4+OoszMrJfq/rAeSbPI+kqZnutQr4PX9sEymj/0wWJmZg1W12Sh7CEwnwcOyvViCtlt899S9pCbPcm6kn6wN+t49dVX6ejo4JVXXulzvNuqwYMHM3r0aLbfvmxHmWa2DapZspB0A1mf+iOVPbHri8A5ZB2l3Zk6Xv1pRJwWEY9KWkjWl0wX8OmiK6Eq6ejoYOedd2bs2LG8tnNXq0ZEsGnTJjo6Oth7770bHY6ZNYlaXg310TKjr+5m+guBC/u63ldeecWJog8kMWLECHzxgJnlDcjuPpwo+sb1Z2alBmSyMDOz/uVkUSdPPfUU++/fl2e61Ma0adN8Y6OZFar7pbPWf7q6uhg0yP9CsyLTZxxB5/qNZctG7T6Sxbd/v84RtR5/09TIJZdcwoIFCwCYM2cORx99NF1dXcyaNYuHH36Yt7/97Vx33XUMGTKEuXPnsmjRIgYNGsShhx7KxRdfzIYNGzjttNNYvTp7zPD8+fOZOnUq559/PmvXruWpp55i5MiR/OpXv2LBggXst1/26Odp06bxpS99iXe84x185jOf4Re/+AVdXV2cf/75zJw5k5dffplTTjmFxx57jHHjxvHyyy83rI7M6qVz/UbaT5tftmz5FWfVNZZWNaCTxVm3n8WyXy/r12VOeNME5s+Y3+00S5cu5etf/zoPPPAAEcGUKVM46KCDWLlyJVdffTVTp05l9uzZfPWrX2X27Nnceuut/PKXv0QSmzdvBuDMM8/k7LPP5oADDmD16tUcdthhrFixYuvy77vvPnbaaScuvfRSFi5cyAUXXEBnZydr167l3e9+N+eeey4HH3wwCxYsYPPmzUyePJlDDjmEK6+8kiFDhrB8+XKWL1/OxIkT+7V+zGxg8jmLGrjvvvs45phjGDp0KMOGDePYY4/lJz/5CWPGjGHq1KkAnHTSSdx3333ssssuDB48mDlz5nDLLbcwZMgQAO666y7OOOMMJkyYwFFHHcVzzz3H888/D8BRRx3FTjvtBMAJJ5zAzTdnj3deuHAhxx9/PAB33HEH8+bNY8KECUybNo1XXnmF1atXc++993LSSdljs9vb22lvb69r3ZhZaxrQRxZFRwC1Uqnb99JLUiUxaNAgHnzwQRYvXsyNN97I5Zdfzt13383vf/977r///q1JIW/o0KFbh/faay9GjBjB8uXLuemmm7jyyiu3xvCd73yHffd9fRfOvjTWzHrKRxY1cOCBB/Ld736Xl156iRdffJFbb72VD3zgA6xevZr77896Xb/hhhs44IADeOGFF3j22Wc5/PDDmT9/PsuWLQPg0EMP5fLLL9+6zC3jyznxxBO56KKLePbZZ3nnO98JwGGHHcZXvvKVrYnr4Ycf3hrb9ddn/Tc+8sgjLF++vL8338wGICeLGpg4cSInn3wykydPZsqUKcyZM4fhw4czbtw4rr32Wtrb2/nNb37D6aefzvPPP8+RRx5Je3s7Bx10EJdeeikAl112GUuWLKG9vZ3x48dzxRVXVFzfcccdx4033sgJJ5ywddx5553Hq6++Snt7O/vvvz/nnXceAKeffjovvPAC7e3tXHTRRUyePLm2lWFmA0JLP4N70qRJUXqPwIoVKxg3blyDIho4XI/WzHr68KPxE6d0ezXUYw890D+BtQhJSyNiUk/m8ZGFmZkVcrIwM7NCAzJZtHLTWjNw/ZlZqQF36ezgwYPZtGkTI0aM8CWivbDleRaDBw9udChmdbFm9SrGT5xSsdzdgWQGXLIYPXo0HR0dfh5DH2x5Up7ZtqArVPHkN7g7kC0GXLLYfvvt/YQ3M7N+NiDPWZiZWf9ysjAzs0JOFmZmVsjJwszMCjlZmJlZIScLMzMr5GRhZmaFBtx9FmZm/am7O7y3pbu7nSzMzLrR3R3e29Ld3W6GMjOzQjVLFpIWSFov6ZHcuN0k3Snp8fR3eK7sHElPSFop6bBaxWVmZj1XyyOLa4AZJePmAosjYh9gcXqPpPHAicB+aZ6vStquhrGZmVkP1CxZRMS9wG9KRs8Erk3D1wJH58bfGBG/jYgngScAPxzazKxJ1PucxR4R0QmQ/u6exu8FrMlN15HGvY6kUyUtkbTE3ZCbmdVHs5zgLveUorKPa4uIqyJiUkRMamtrq3FYZmYG9U8W6ySNAkh/16fxHcCY3HSjgbV1js3MzCqod7JYBMxKw7OA23LjT5S0o6S9gX2AB+scm5mZVVCzm/Ik3QBMA0ZK6gC+CMwDFkr6JLAaOB4gIh6VtBB4DOgCPh0Rv6tVbGZm1jM1SxYR8dEKRdMrTH8hcGGt4jEzs95rlhPcZmbWxJwszMyskJOFmZkVcrIwM7NCThZmZlbIycLMzAo5WZiZWSEnCzMzK+RkYWZmhZwszMyskJOFmZkVcrIwM7NCNetI0MysXqbPOILO9Rsrlq/peJr2OsYzEDlZmFnL61y/kfbT5lcsf/KcY+sXzADlZigzMyvkZGFmZoWcLMzMrJCThZmZFXKyMDOzQk4WZmZWyMnCzMwKOVmYmVkhJwszMyvkO7jNrOmVduexauoKAMZfNgVwdx714GRhZk2vtDuPjZs+DbB1nLvzqD03Q5mZWSEnCzMzK9SQZCHpbEmPSnpE0g2SBkvaTdKdkh5Pf4c3IjYzM3u9uicLSXsBfwlMioj9ge2AE4G5wOKI2AdYnN6bmVkTaFQz1CBgJ0mDgCHAWmAmcG0qvxY4ujGhmZlZqboni4h4GrgYWA10As9GxB3AHhHRmabpBHYvN7+kUyUtkbRkw4YN9QrbzGyb1ohmqOFkRxF7A3sCQyWdVO38EXFVREyKiEltbW21CtPMzHIa0Qx1CPBkRGyIiFeBW4D3A+skjQJIf9c3IDYzMyujEcliNfBeSUMkCZgOrAAWAbPSNLOA2xoQm5mZlVH3O7gj4gFJ3wYeArqAh4GrgGHAQkmfJEsox9c7NjMzK68h3X1ExBeBL5aM/i3ZUYaZmTUZ38FtZmaF3JGgmTWF0p5l89yrbOM5WZhZUyjtWTbPvco2npuhzMyskJOFmZkVcrIwM7NCThZmZlbIycLMzAo5WZiZWSEnCzMzK+RkYWZmhZwszMyskJOFmZkVcrIwM7NCThZmZlbIycLMzAo5WZiZWaGqkoWkqdWMMzOzganaI4uvVDnOzMwGoG4ffiTpfcD7gTZJf5Ur2gXYrpaBmZk1uzWrVzF+4pSK5aN2H8ni279fx4hqp+hJeTsAw9J0O+fGPwccV6ugzMxaQVeo4tP9AJZfcVbdYqm1bpNFRPwY+LGkayJiVZ1iMjOzJlPtM7h3lHQVMDY/T0QcXIugzMysuVSbLG4GrgD+Dfhd7cIxs4Fq+owj6Fy/sWL5mo6naa9jPNYz1SaLroj4Wk0jMbMBrXP9xm7b958859j6BWM9Vu2ls9+T9ClJoyTttuVV08jMzKxpVHtkMSv9/VxuXABv7d9wzMysGVWVLCJi7/5cqaRdyc5/7E+WdGYDK4GbyE6iPwWcEBHP9Od6zcysd6pKFpI+UW58RFzXy/V+Gbg9Io6TtAMwBDgXWBwR8yTNBeYCn+/l8s3MrB9V2wz1ntzwYGA68BDQ42QhaRfgQOBkgIj4X+B/Jc0EpqXJrgXuwcnCzKwpVNsM9Zn8e0lvBL7Ry3W+FdgAfF3SHwNLgTOBPSKiM62vU9Lu5WaWdCpwKsCb3/zmXoZgZmY90dsuyl8C9unlvIOAicDXIuJdwItkTU5ViYirImJSRExqa2vrZQhmZtYT1Z6z+B7ZiWjIOhAcByzs5To7gI6IeCC9/zZZslgnaVQ6qhgFrO/l8s3MrJ9Ve87i4txwF7AqIjp6s8KI+LWkNZL2jYiVZOc/HkuvWcC89Pe23izfzMz6X7XnLH4saQ/+cKL78T6u9zPA9elKqP8BTiFrElso6ZPAauD4Pq7DzMz6SbXNUCcA/0x2hZKAr0j6XER8uzcrjYhlwKQyRdN7szwzM6utapuhvgC8JyLWA0hqA+4iO99gZmYDXLVXQ71hS6JINvVgXjMza3HVHlncLumHwA3p/UeAH9QmJDMzazZFz+B+G9nNcp+TdCxwANk5i/uB6+sQn5mZNYGipqT5wPMAEXFLRPxVRJxNdlQxv7ahmZlZsyhKFmMjYnnpyIhYQtY7rJmZbQOKksXgbsp26s9AzMyseRUli59J+vPSkenGuaW1CcnMzJpN0dVQZwG3SvoYf0gOk4AdgGNqGJeZmTWRbpNFRKwD3i/pT8ieagfw/Yi4u+aRmZlZ06i2b6gfAT+qcSxmZtakfBe2mZkVcrIwM7NCThZmZlbIycLMzAo5WZiZWSEnCzMzK+RkYWZmhZwszMysULUPPzIzKzR9xhF0rt9YtmxNx9O01zke6z9OFmbWbzrXb6T9tPlly54859j6BmP9ys1QZmZWyMnCzMwKOVmYmVkhJwszMyvkZGFmZoWcLMzMrFDDkoWk7SQ9LOnf0/vdJN0p6fH0d3ijYjMzs9dq5JHFmcCK3Pu5wOKI2AdYnN6bmVkTaEiykDQaOAL4t9zomcC1afha4Og6h2VmZhU06shiPvA3wO9z4/aIiE6A9Hf3cjNKOlXSEklLNmzYUPNAzcysAclC0pHA+ohY2pv5I+KqiJgUEZPa2tr6OTozMyunEX1DTQWOknQ4MBjYRdI3gXWSRkVEp6RRwPoGxGZmZmXU/cgiIs6JiNERMRY4Ebg7Ik4CFgGz0mSzgNvqHZuZmZXXTPdZzAM+KOlx4IPpvZmZNYGGdlEeEfcA96ThTcD0RsZjZmblNdORhZmZNSknCzMzK+RkYWZmhfxYVTOrWnfP2AY/Z3sgc7Iws6p194xt8HO2BzI3Q5mZWSEnCzMzK+RkYWZmhZwszMyskJOFmZkVcrIwM7NCThZmZlbIycLMzAo5WZiZWSEnCzMzK+RkYWZmhZwszMyskJOFmZkVcrIwM7NCThZmZlbIycLMzAo5WZiZWSEnCzMzK+RkYWZmhfwMbrNtzPQZR9C5fmPZslG7j2Tx7d+vc0TWCpwszLYxnes30n7a/LJly684q66xWOtwM5SZmRWqe7KQNEbSjyStkPSopDPT+N0k3Snp8fR3eL1jMzOz8hrRDNUFfDYiHpK0M7BU0p3AycDiiJgnaS4wF/h8A+IzM+sXa1avYvzEKRXLW+kcUd2TRUR0Ap1p+HlJK4C9gJnAtDTZtcA9OFmYWQvrClU8PwStdY6ooecsJI0F3gU8AOyREsmWhLJ7hXlOlbRE0pINGzbULVYzs21Zw5KFpGHAd4CzIuK5aueLiKsiYlJETGpra6tdgGZmtlVDkoWk7ckSxfURcUsavU7SqFQ+CljfiNjMzOz1GnE1lICrgRURcUmuaBEwKw3PAm6rd2xmZlZeI66Gmgp8HPiFpGVp3LnAPGChpE8Cq4HjGxCbmZmV0Yiroe4DVKF4ej1jMTOz6vgObjMzK+S+ocxsq6KbyNZ0PE17HeOx5uFkYWZbFd1E9uQ5x9YvGGsqboYyM7NCThZmZlbIzVBmA0x3DzcCn3ew3nGyMBtgunu4Efi8g/WOm6HMzKyQk4WZmRVysjAzs0JOFmZmVsjJwszMCjlZmJlZIScLMzMr5PsszFpQdzfe+aY7qwUnC7MW1N2Nd77pzmrBzVBmZlbIRxZmvVTUB9Oo3Uey+Pbv1zEis9pxsjDrpaI+mJZfcVbdYjGrNTdDmZlZIScLMzMr5GYosxrp7nnWRecz/EwKazZOFmY10t3zrIvOZ/iZFNZsnCxsm+Yrmsyq42RhTa+WX+hFv+C/f+4xFZuS+tIU1F0TVV+XbVYLThbW9Bp5iWp3TUl9aQrqbrl9XbZZLThZWMvry4lks0ZqpX236ZKFpBnAl4HtgH+LiHkNDsmaXF9OJJs1Uivtu02VLCRtB/wL8EGgA/iZpEUR8Vgt1tddW3hfsnqrnjTtS9xF867rXMseo/bscRn43IBtm4r23Xp/jzRVsgAmA09ExP8ASLoRmAnUJFl01xbel6zeqt1A9CXuonlvO+fYiuXdlYHPDdi2qWjfrff3iCKirivsjqTjgBkRMSe9/zgwJSLOyE1zKnBqersvsLLugfbNSKDyT/Dm1qqxt2rc4NgbpVVjrzbut0REW08W3GxHFioz7jXZLCKuAq6qTzj9T9KSiJjU6Dh6o1Vjb9W4wbE3SqvGXsu4m61vqA5gTO79aGBtg2IxM7Ok2ZLFz4B9JO0taQfgRGBRg2MyM9vmNVUzVER0SToD+CHZpbMLIuLRBofV31q2CY3Wjb1V4wbH3iitGnvN4m6qE9xmZtacmq0ZyszMmpCThZmZFXKyqAFJu0m6U9Lj6e/wMtMMlvSgpJ9LelTSBbmy8yU9LWlZeh3eInEXzt/g2MdI+pGkFSn2M3NlDanzfoq9qes9TbdA0npJj5SMb9p9vSDuVqjzGZJWSnpC0tzc+F7VuZNFbcwFFkfEPsDi9L7Ub4GDI+KPgQnADEnvzZVfGhET0usHNY8409e4q5m/VqpZdxfw2YgYB7wX+LSk8bnyRtQ59D32Zq93gGuAGRXKmnVfh8pxN3Wd57pO+hAwHvhon/f1iPCrn19kd5WPSsOjgJUF0w8BHiK7Wx3gfOCvWzDuHs3fyNjTdLcBH2xknfdT7C1R78BY4JGScU2/r1eIu6nrHHgf8MPc+3OAc/pS5z6yqI09IqITIP3dvdxEkraTtAxYD9wZEQ/kis+QtDwdBtfrELevcVc1f430aN2SxgLvAhpd59D32Fum3ito2n29hvP3RTXr3gtYk3vfkcZt0eM6b6r7LFqJpLuAN5Up+kK1y4iI3wETJO0K3Cpp/4h4BPga8A9kXZ38A/AlYHafg6bmcddUf8SeljMM+A5wVkQ8l0bXrM7TOmsZe031V+wVNPW+3ij9EHt3XSf1qs6dLHopIg6pVCZpnaRREdEpaRTZL/DulrVZ0j1kbaOPRMS63LL+Ffj3fgq7pnEDPZq/EbFL2p7sy/b6iLglt+ya1Xlafs1ipwXqvZtlt8S+Xkaz13nFrpN6W+duhqqNRcCsNDyLrH35NSS1pV/mSNoJOAT4ZXo/KjfpMWRfxPXQp7irmb+GqoldwNXAioi4pKSsUXUOfYy9mvlrqE/rbuZ9vcbz90U1667YdVKv67xeJ2W2pRcwguwqhcfT393S+D2BH6ThduBhYHn6Z/1dbv5vAL9IZYtIJ7NaIO6y8zdR7AeQHXovB5al1+GNrPN+ir2p6z29vwHoBF4l+9X7yWbf1wviboU6Pxz4b+BXwBdy43tV5+7uw8zMCrkZyszMCjlZmJlZIScLMzMr5GRhZmaFnCzMzKyQk4WZmRVysjAzs0L/B6GFJbOsepsxAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr_all_pests_best_estimator = lr_all_pests_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "lr_cv_scores_all_pests = model_selection.cross_val_score(\n",
    "                    lr_all_pests_best_estimator,\n",
    "                    X_train_all_pests,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "lr_train_pred_all_pests = lr_all_pests_best_estimator.predict(X_train_all_pests)\n",
    "sns.histplot(lr_train_pred_all_pests).set(title='Predicted vs observed - training data - all pests')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6381786b",
   "metadata": {},
   "source": [
    "#### LRR: with features - metrics - predictor: pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "370c395a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.062\n",
      "Train MAE:  0.045\n",
      "Test MAE:  0.046\n",
      "Train R-squared:  0.125\n",
      "Test R-squared:  0.095\n",
      "Train RMSE:  0.095\n",
      "Test RMSE:  0.079\n"
     ]
    }
   ],
   "source": [
    "lr_all_pests_test_pred = lr_all_pests_best_estimator.predict(X_test_all_pests)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(lr_cv_scores_all_pests): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=lr_train_pred_all_pests): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=lr_all_pests_test_pred): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=lr_train_pred_all_pests): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=lr_all_pests_test_pred): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=lr_train_pred_all_pests, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=lr_all_pests_test_pred, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f056642e",
   "metadata": {},
   "source": [
    "#### LRR: with features - plotting observed vs predicted - predictor: pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "0fe1448d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))\n",
    "\n",
    "ax1.scatter(y_test, lr_all_pests_test_pred, c=\"orange\")\n",
    "\n",
    "p1 = max(max(lr_all_pests_test_pred), max(y_test))\n",
    "p2 = min(min(lr_all_pests_test_pred), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Linear Ridge regression on pests - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(lr_all_pests_test_pred), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title='Linear ridge regression on pests - natural scale')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14b6571a",
   "metadata": {},
   "source": [
    "#### LRR: with features - performance - predictor: pests\n",
    "\n",
    "All metrics for linear ridge regression with pest data are better than the baseline, so this is a better model than the baseline. The difference between train and test data indicate that the model may be slightly underfit (e.g. the test RMSE is lower than the training RMSE. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d87ccc03",
   "metadata": {},
   "source": [
    "#### LRR: with features - fit model - predictor: varroa mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "id": "833053d9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 40 candidates, totalling 200 fits\n",
      "Elapsed time: 0.74 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "lr_mites_regions = lr_grid_w_features.fit(X_train_mites_regions, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d7aea937",
   "metadata": {},
   "source": [
    "#### LRR: with features - CV score distribution - predictor: mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "id": "c025f63c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.03 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f94824fb640>"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr_mites_regions_best_estimator = lr_mites_regions.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "lr_cv_scores_mites_regions = model_selection.cross_val_score(\n",
    "                    lr_mites_regions_best_estimator,\n",
    "                    X_train_mites_regions,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "lr_train_pred_mites_regions = lr_mites_regions_best_estimator.predict(X_train_mites_regions)\n",
    "sns.histplot(lr_train_pred_mites_regions).set(title='Predicted vs observed - training data - mites and regions')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e5e60640",
   "metadata": {},
   "source": [
    "#### LRR: with features - metrics - predictor: mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "id": "37456385",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.065\n",
      "Train MAE:  0.048\n",
      "Test MAE:  0.050\n",
      "Train R-squared:  0.054\n",
      "Test R-squared:  0.042\n",
      "Train RMSE:  0.099\n",
      "Test RMSE:  0.081\n"
     ]
    }
   ],
   "source": [
    "lr_test_pred_mites_regions = lr_mites_regions_best_estimator.predict(X_test_mites_regions)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(lr_cv_scores_mites_regions): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=lr_train_pred_mites_regions): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=lr_test_pred_mites_regions): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=lr_train_pred_mites_regions): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=lr_test_pred_mites_regions): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=lr_train_pred_mites_regions, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=lr_test_pred_mites_regions, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "909cd682",
   "metadata": {},
   "source": [
    "#### LRR: with features - plotting observed vs predicted - predictor: mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "id": "07888ddc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 4))\n",
    "\n",
    "ax1.scatter(y_test, lr_test_pred_mites_regions, c=\"orange\")\n",
    "\n",
    "p1 = max(max(lr_test_pred_mites_regions), max(y_test))\n",
    "p2 = min(min(lr_test_pred_mites_regions), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Linear Ridge regression on mites and regions - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(lr_test_pred_mites_regions), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title=\"Linear Ridge regression on mites and regions - natural scale\")\n",
    "ax2.axis('equal')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f064ed2a",
   "metadata": {},
   "source": [
    "#### LRR: with features - performance - predictor: mites and region data\n",
    "\n",
    "Linear ridge regression on the varroa mite and region data together has mixed performance, as compared to the baseline. The model was originally optimized for mean absolute error, and performs about the same as the naive model on that metric. However, the R-squared is considerably better than the naive model, and the RMSE is marginally better. Test and train results for MAE are similar, indicating that the model is an accurate predictor."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec5ecc58",
   "metadata": {},
   "source": [
    "#### LRR: with features - fit model - predictor: Patsy matrix of mites, pesticides, and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "id": "a0f5f4a9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 40 candidates, totalling 200 fits\n",
      "Elapsed time: 0.32 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "lr_matrix_mod = lr_grid_w_features.fit(train_model_matrix, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0db458be",
   "metadata": {},
   "source": [
    "#### LRR: with features - CV score distribution - predictor: Patsy matrix of mites, pesticides, and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "id": "aff94adb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.01 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f9438bb3190>"
      ]
     },
     "execution_count": 132,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr_matrix_best_estimator = lr_matrix_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "lr_cv_scores_matrix = model_selection.cross_val_score(\n",
    "                    lr_matrix_best_estimator,\n",
    "                    train_model_matrix,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "lr_train_pred_matrix = lr_matrix_best_estimator.predict(train_model_matrix)\n",
    "sns.histplot(lr_train_pred_matrix, label='CV scores').set(title='Predicted vs observed - training data - Patsy matrix')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b2c09a7",
   "metadata": {},
   "source": [
    "#### LRR: with features - metrics - predictor: Patsy matrix of mites, pesticides, and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "id": "ac66efee",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.065\n",
      "Train MAE:  0.049\n",
      "Test MAE:  0.049\n",
      "Train R-squared:  0.058\n",
      "Test R-squared:  0.042\n",
      "Train RMSE:  0.099\n",
      "Test RMSE:  0.081\n"
     ]
    }
   ],
   "source": [
    "lr_test_pred_matrix = lr_matrix_best_estimator.predict(test_model_matrix)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(lr_cv_scores_matrix): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=lr_train_pred_matrix): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=lr_test_pred_matrix): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=lr_train_pred_matrix): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=lr_test_pred_matrix): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=lr_train_pred_matrix, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=lr_test_pred_matrix, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cdbcf479",
   "metadata": {},
   "source": [
    "#### LRR: with features - plotting observed vs predicted - predictor: Patsy matrix of mites, pesticides, and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "id": "d38610f3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 4))\n",
    "\n",
    "ax1.scatter(y_test, lr_test_pred_matrix, c=\"orange\")\n",
    "\n",
    "p1 = max(max(lr_test_pred_matrix), max(y_test))\n",
    "p2 = min(min(lr_test_pred_matrix), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Linear Ridge on Patsy matrix - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(lr_test_pred_matrix), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title=\"Linear Ridge on Patsy matrix - natural scale\")\n",
    "ax2.axis('equal')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d2503772",
   "metadata": {},
   "source": [
    "#### LRR: with features - performance - predictor: Patsy matrix of mites, pesticides, and regions\n",
    "\n",
    "All metrics for linear ridge regression with pest data are better than the baseline, so this is a better model than the baseline. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0340c622",
   "metadata": {},
   "source": [
    "### Linear Ridge regression - overall conclusions\n",
    "\n",
    "The two best linear ridge regression models that provide value in their predictive abilities are the model of all the pest data, and the model using the Patsy matrix of varroa mites, pesticides, and region data. Of the two, the model using all the pest data as the predictor is the better model in terms of metric values. However, the Patsy model of varroa mite, pesticide, and region data has closer alignment between the training and testing sets, indicating that the consistency of the Patsy matrix as predictor may be higher."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "822a9f38",
   "metadata": {},
   "source": [
    "## Random Forest regression"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f9cff82c",
   "metadata": {},
   "source": [
    "#### RF - create model object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "id": "99097944",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_base_model = ensemble.RandomForestRegressor(random_state=42, bootstrap = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5fac24e8",
   "metadata": {},
   "source": [
    "#### RF - create params object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "id": "d7ae6c97",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_params = {'model__n_estimators': [50,100,150], \n",
    "             'model__max_depth': [None,4,8], \n",
    "             'model__min_samples_split': np.arange(2,20,5),\n",
    "             'model__max_features': [0.3,0.5,1,'sqrt','auto'],\n",
    "             'model__max_samples': [20]}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "78a512dd",
   "metadata": {},
   "source": [
    "#### RF - create pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "id": "0400c794",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_pipeline = pipeline.Pipeline(steps=[(\"model\", rf_base_model)])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "36809fad",
   "metadata": {},
   "source": [
    "#### RF - create baseline grid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "id": "38cea558",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_baseline_grid = model_selection.GridSearchCV(\n",
    "                        estimator=rf_pipeline,\n",
    "                        param_grid=rf_params,\n",
    "                        refit=True,\n",
    "                        scoring='neg_mean_absolute_error',\n",
    "                        cv=cross_validator,\n",
    "                        verbose=1\n",
    "                    )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7b1e5ae3",
   "metadata": {},
   "source": [
    "### Random Forest - naive model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "493438bf",
   "metadata": {},
   "source": [
    "#### RF: naive - fit model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "id": "ee85f08e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 180 candidates, totalling 900 fits\n",
      "Elapsed time: 79.58 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "rf_base_mod = rf_baseline_grid.fit(X_train_null, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c27274e5",
   "metadata": {},
   "source": [
    "#### RF: naive - CV score distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "id": "7bef4f08",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.54 seconds\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEICAYAAACwDehOAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAAAeMklEQVR4nO3debhd893+8fctQRJRkgiNBIkKgieGBgmq0VBiSmpMi4ZHaZ9q6VyhA620qh4/Hq0hpYZSpAkVSg0hbQ0NMTQkoVJTjoxUZBAq8fn9sb5nWU7OsJOcdfaJc7+ua19nzevzXXu49xrO2ooIzMzMANapdgFmZtZ6OBTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUCiRpHMk3VDtOtZ2ko6TdG+161gVkk6U9FCF014r6byya6pE3bolLZG0dQusd5KkL5W9nkbW/7Kk/au1/tVR1jZrc6GQnvxl6cU+N70hO1e7rjUhabCk91Obah93tOD6e0sKSe3LWH5E3BgRny1j2da4iOgcES82Nk3Zz7+1rDYXCslhEdEZ2AXYFRhV3XKaxez0Bq59HLaqC5DUrozCrDqUaavv8VZnbQnNNv2CiYi5wD1k4QCApDMl/UvSYknTJX2uMO5ESQ9JulDSm5JekjS0ML6PpL+kee8DNimuT9LhkqZJWph2/foVxr0s6buSpkpaKulqSZtJujst735JXVa1jZL6pXUtTOs+vDDuWkmXS7pL0lJgP0mbSxovaUFq3+mF6feQNEXSIknzJF2URv01/V2Y9lIG1VPHOZLGSro+tWeapAGrst1T9xWSLqyz7NslfSt1N1h/PTVdK+mytI2XSHpY0sclXZye3+ck7VrhtuwmaULaNo8Bn6izru0l3Sfp35Kel3RMQ3U1Jm2LhyVdKumtVOOQwvhJkkZLehh4G9i6sXVXUHdI2iZ1d5T0v5JeSet+SFJHGnj+Jf23pBlpW94jaavCcg9Itb8l6VeAGmjv5sr27LsWhu0q6XVJ60raRtl77q007JZGtl2D779k9/Tae1PSNZI6pPk2kXRnmu/fkv6mFLaNvd7Sa36cpBskLQLOaqwtzbXN1lhEtKkH8DKwf+ruBTwDXFIYfzSwOVlgHgssBXqkcScC7wGnAO2A/wFmA0rjHwUuAtYH9gUWAzekcdumZR0ArAt8D5gJrFeo6+/AZkBPYD7wJNmezPrAA8CPG2jTYKCmnuHrpnWcBawHfCbVtF0afy3wFrB3am8n4AngR2n6rYEXgQML7TshdXcGBqbu3kAA7RvZ7ucA7wAHp233c+Dvq7DdH0rd+wKzCtu8C7CsMG+D9ddT07XA68AngQ5pG78EfDHVeB7wYIXb8mZgLLABsBPwWqHmDVLNJwHtgd3Sencs1HFeoa6FwD4N1HwisBz4Zqrp2PQcdk3jJwGvAjumdW3UxLobrDuND2Cb1P3rtPyeafvsRfbaXOn5B4an7dUvrfcHwCNp3CbAIuCo1IZvpjZ9qYE2PwCcUuj/JXBF6r4JODs99x0a2W6VvP+eBbYAugIP1z4nZK/VK9J86wKfIvtAbvT1Rvaafy9ti3WAjk20pdm22Rp9RlbzA7oaj/TkLyF7QwcwEdi4kemfBoYV3pAzC+M6pWV8HNgyPUkbFMb/ng9C4YfA2MK4dcjegIMLdR1XGD8euLzQ/3Xgjw3UOBh4n+zDpPZxTHrxzgXWKUx7E3BO6r4WuL4wbk/g1TrLHgVck7r/CpwLbFJnmt5UFgr3F/p3AJatwnav/YAV2Yfevqn/FOCBSuqvZx3XAr+ps41nFPr/C1iYuhvclmQfkO8B2xfG/axQ87HA3+qs+0pSyFMnFJp4/Z5I4YtIGvYYH4T1JOAnhXENrrupulN/ANuQvV6XATvXU9NKzz9wN3Byndf728BWZKFb/EIgoIaGQ+FLhedYZCFX+/xfD4wBejWx3Sp5/32lMP5g4F+p+yfA7aRwXIX3yznAX1ehLc22zdbk0VYPHw2PiA3JPky3p3CYR9IXJT2ddhUXkn17Kh4GmlvbERFvp87OZN9U34yIpYVpXyl0b17sj4j3yV4QPQvTzCt0L6unv7ET4rMjYuPCY2xa56y0rmJNxXXOKnRvBWxe2/bU/rPI9l4ATib7xvWcpMclHdpIPfWZW+h+G+igdJy1gu0OQGTviJuBz6dBXwBurLD++lS6zRvblt3JvtnNqjOu1lbAnnXqOo7sy8TqeC1th+K6Ni/0131OG1p3U3UXbUL2TfxfFda4FXBJYZ3/Jvsg65lqzdeZ2jKrvoUk44BBkjYn21MM4G9p3PfSch9Lh4b+u4FlVPL+q7sdarfpL8m+wd8r6UVJZxba2NTrrW67GmtLc26z1bZWnPgoS0T8RdK1wIXA8HT87jfAEODRiFgh6WkqO3Y3B+giaYNCMGxJ9qRD9u3uv2onliSyXdXXmqMtDZgNbCFpncKH2ZbAPwvTFD9cZgEvRUTf+hYWES8An0/HU48AxknqVmcZq2w1tvtNZG/Q88m+rdWef2i0/jXU2LZcQLaXuAXwXGFcrVnAXyLigGaqpackFYJhS2BCYXzd57TedSu7sKCxuoteJzv89wngH3XG1ff8zwJGR8SNdUdI6pvWWduvYn9dEbFQ2SXJx5AdWrmptu2RnRc8JS1nH+B+SX+NiJl1FlPJ+69Yw5ZpHiJiMfBt4NuSdgQelPQ4lb3ePrRtGmsLzbjN1kRb3VMouhg4QNIuZMdVg+xNjqSTyL6xNikiXgGmAOdKWi+9QItXAI0FDpE0JJ1U+jbwLvBIM7WjPpPJjqN+L52UG5xqurmB6R8DFkn6vrKTiu0k7SRpdwBJx0vqnj4UF6Z5VpBtr/fJjqmujlXa7hHxVJr2KuCeiKitpdH611CD2zIiVgC3AudI6iRpB2BkYd47gW0lnZDmXVfS7lr5RGelNgVOT8s5muzD5a4Gpm1w3RXUnUvP+W+Bi5SdXG0naZCk9an/+b8CGJU+RJG0UaoV4E/AjpKOSHuKp9P0XtPvyQ6hHJm6Scs9WlKv1Psm2etoRT3zV/L+O01SL2Ungs8CbknrOFTZCW2RHddfkR6r+3qrty00/zZbLW0+FCJiAdlxyR9GxHTgf8lOqM4j+2bx8Cos7gtk31z/TXbM9vrCep4HjgcuJfvWdRjZpbH/aYZm1Cst+3BgaFrnZcAXI+K5BqZfkeraheyE6+tkH7wbpUkOAqZJWgJcAoyIiHfSYbTRwMNp13fgKta5Otv9JmB/Cm+qCupfbRVsy6+RHWqaS3aO4JrCvIuBzwIjyL59zgV+QXaSdiXKruD5VCPlTAb6pjpGA0dFxBsN1N3Uuhusux7fIbsw43Gy1/gvyM6xrPT8R8RtafzNyq68eZZs2xERr5NdWHA+8EZqS1PP94Q03byIKO6p7A5MTq/JCcAZEfFSPduhkvff74F7yU4Wv0h2oQFpvfeTnYt8FLgsIiatweut3raUsM1WS+0VHGa2FpB0ItnJxX2qXYt9NLX5PQUzM/uAQ8HMzHI+fGRmZjnvKZiZWW6t/j+FTTbZJHr37l3tMsxW8vwbzwOwXbftqlyJ2cqeeOKJ1yOie33j1upQ6N27N1OmTKl2GWYrGXztYAAmnTipqnWY1UdSQ/+57sNHZmb2AYeCmZnlHApmZpZbq88pmNlH13vvvUdNTQ3vvPNOtUtZa3Xo0IFevXqx7rrrVjyPQ8HMWqWamho23HBDevfuTXYvOlsVEcEbb7xBTU0Nffr0qXg+Hz4ys1bpnXfeoVu3bg6E1SSJbt26rfKelkPBzFotB8KaWZ3t51AwM7OcQ8HMbBW8/PLL7LRTRb+91aIGDx7cLP/M26ZPNPfcYktm15TyM6fW1p2Y/dFJPvyxuu6++26WLl3a9IQtbPbs2SxbtqxZ76awfPly2rdftY/jdddbj53792+2Gmq16VCYXTOLY68s89cwra164I3TAPjMlb+uciVrr05dO9Jlq22qXQZXX34p4276HQDHHDeSA4YeCuu042e/vIjpz0yl9ye24cJfjaFjp05c8NMfMfGeu2jfrj37DP4Mo879GW+8voAfffcbzH6tBoAf/PR8PrnnIC654GfMnzeHmldfpWu3brz68kv8/OLL2Hb77FdavzB8KGed+zO23mZbfnLWd3h+xnSWL1/O6d8dxQFDD2XO808zYsQIpk+fTr9+/Vi2bFmztLdNh4KZrR3Oe+j7TH99arMuc4dN+vODfX7R6DTP/uMpxt98A+PvfpCI4Mih+7HHXvvw4swX+Pn/+zWf3HMQZ57xP9x4zW846gsncN9dd3DvI08iiUVvLcxq/8H3OenLpzFg4F7MrpnFSccO556Hn0jLf5pb7riXDh078tsrfsVdt9/Kttufzfx5c5k/dw477bwrF44+h4H7fJrzL7mcRW8t5IgDB7P3vvsxfvx4OnXqxNSpU5k6dSq77bZbs2wXn1MwM2vAlMmP8tmhh9Fpgw3YoHNnDjzkcKb8/RF69OzFJ/ccBMCwo0YwZfKjdN7wY6y/fgdGffM07rnzdjp07ATAw399kHNHfYfD9tuLL59wLEsWL2bJksUADDnwYDp07AjAwcOO4O47bgPgrttvZejhnwPgoUkPcOWlF3HYfntx3PCDeffdd5n92iyeeuopjj/+eAD69+9P/2Y6lOQ9BTNr9Zr6Rl+Whn6ErO6lnpJo37494++ZxKN/m8Sdt43nd78dww23/ol4/33+cNfE/MO/qFOnTnn3x3tsTpcuXXlu2rP86Y/jOe/C/8tr+PVvb2DrbbZtso7m4D0FM7MG7D5ob+67+06Wvf02by9dyr133ZEfBnry8ckA3HHbH/jknoNYumQJSxYtYvD+B/KD885nxrPPALDP4CH87uor82VOf6bhw2CHDD+KMb+6mMWLF7HdDjsC8Kn9hnD9VVfmATXtmX8AsOuuu3LjjTcC8OyzzzJ1avMcXnMomJk1YKf+u3DEiOM44qDBHDl0P445biQbbbQxn9h2O2675fcc8umBvPXmmxx34pdYunQJpxx/NId8eiBfGD6Us3/ycwB+OPoCnvnHUxzy6YEcuM8Abrru6gbXN/Sw4fzpj+M4+PAj8mFf+9b3Wf7eexwyeCBD992Di8//KQBHHnkkS5YsoX///lxwwQXssccezdLmtfo3mgcMGBBrclmYJF99ZKXIrz7q5quPVtdXd+lIzz7Vv/qotXrzlecYMGBAk9PNmDGDfv36fWiYpCciot6ZvadgZmY5h4KZmeUcCmbWKgUNX/1jlVmd7edQMLNWacHb7/Pu4rccDKup9vcUOnTosErz+f8UzKxVuuvFd4H5dO/0Or6D1Mre/vfrzJgxo9Fpan95bVU4FMysVXp7OYz757vVLqPVuuXLQ0vZi/LhIzMzyzkUzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHKlhoKkb0qaJulZSTdJ6iCpq6T7JL2Q/nYpTD9K0kxJz0s6sMzazMxsZaWFgqSewOnAgIjYCWgHjADOBCZGRF9gYupH0g5p/I7AQcBlktqVVZ+Zma2s7MNH7YGOktoDnYDZwDDgujT+OmB46h4G3BwR70bES8BMoHl+SsjMzCpSWihExGvAhcCrwBzgrYi4F9gsIuakaeYAm6ZZegKzCouoScM+RNKpkqZImrJgwYKyyjcza5PKPHzUhezbfx9gc2ADScc3Nks9w1a621NEjImIARExoHv37s1TrJmZAeUePtofeCkiFkTEe8CtwF7APEk9ANLf+Wn6GmCLwvy9yA43mZlZCykzFF4FBkrqJEnAEGAGMAEYmaYZCdyeuicAIyStL6kP0Bd4rMT6zMysjtJ+TyEiJksaBzwJLAeeAsYAnYGxkk4mC46j0/TTJI0FpqfpT4uIFWXVZ2ZmKyv1R3Yi4sfAj+sMfpdsr6G+6UcDo8usyczMGub/aDYzs5xDwczMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs1ypoSBpY0njJD0naYakQZK6SrpP0gvpb5fC9KMkzZT0vKQDy6zNzMxWVvaewiXAnyNie2BnYAZwJjAxIvoCE1M/knYARgA7AgcBl0lqV3J9ZmZWUFooSPoYsC9wNUBE/CciFgLDgOvSZNcBw1P3MODmiHg3Il4CZgJ7lFWfmZmtrMw9ha2BBcA1kp6SdJWkDYDNImIOQPq7aZq+JzCrMH9NGvYhkk6VNEXSlAULFpRYvplZ21NmKLQHdgMuj4hdgaWkQ0UNUD3DYqUBEWMiYkBEDOjevXvzVGpmZkC5oVAD1ETE5NQ/jiwk5knqAZD+zi9Mv0Vh/l7A7BLrMzOzOkoLhYiYC8yStF0aNASYDkwARqZhI4HbU/cEYISk9SX1AfoCj5VVn5mZrax9ycv/OnCjpPWAF4GTyIJorKSTgVeBowEiYpqksWTBsRw4LSJWlFyfmZkVlBoKEfE0MKCeUUMamH40MLrMmszMrGH+j2YzM8s5FMzMLOdQMDOznEPBzMxyDgUzM8s5FMzMLOdQMDOznEPBzMxyDgUzM8tVFAqS9q5kmJmZrd0q3VO4tMJhZma2Fmv03keSBgF7Ad0lfasw6mOAfyrTzOwjpqkb4q0HdE7TbVgYvgg4qqyizMysOhoNhYj4C/AXSddGxCstVJOZmVVJpbfOXl/SGKB3cZ6I+EwZRZmZWXVUGgp/AK4ArgL8wzdmZh9RlYbC8oi4vNRKzMys6iq9JPUOSV+V1ENS19pHqZWZmVmLq3RPYWT6+93CsAC2bt5yzMysmioKhYjoU3YhZmZWfRWFgqQv1jc8Iq5v3nLMzKyaKj18tHuhuwMwBHgScCiYmX2EVHr46OvFfkkbAb8rpSIzM6ua1b119ttA3+YsxMzMqq/Scwp3kF1tBNmN8PoBY8sqyszMqqPScwoXFrqXA69ERE0J9ZiZWRVVdPgo3RjvObI7pXYB/lNmUWZmVh2V/vLaMcBjwNHAMcBkSb51tpnZR0ylh4/OBnaPiPkAkroD9wPjyirMzMxaXqVXH61TGwjJG6swr5mZrSUq3VP4s6R7gJtS/7HAXeWUZGZm1dLUbzRvA2wWEd+VdASwDyDgUeDGFqjPzMxaUFOHgC4GFgNExK0R8a2I+CbZXsLF5ZZmZmYtralQ6B0RU+sOjIgpZD/NaWZmHyFNhUKHRsZ1bM5CzMys+poKhcclnVJ3oKSTgSfKKcnMzKqlqauPvgHcJuk4PgiBAcB6wOcqWYGkdsAU4LWIODT9jOctZIefXgaOiYg307SjgJOBFcDpEXHPqjTGzMzWTKN7ChExLyL2As4l+wB/GTg3IgZFxNwK13EGMKPQfyYwMSL6AhNTP5J2AEYAOwIHAZelQDEzsxZS6b2PHoyIS9PjgUoXLqkXcAhwVWHwMOC61H0dMLww/OaIeDciXgJmAntUui4zM1tzZf9X8sXA94D3C8M2i4g5AOnvpml4T2BWYbqaNOxDJJ0qaYqkKQsWLCilaDOztqq0UJB0KDA/Iio9Ia16hsVKAyLGRMSAiBjQvXv3NarRzMw+rNLbXKyOvYHDJR1MdmnrxyTdAMyT1CMi5kjqAdTeU6kG2KIwfy9gdon1mZlZHaXtKUTEqIjoFRG9yU4gPxARxwMTgJFpspHA7al7AjBC0vqS+pD93OdjZdVnZmYrK3NPoSHnA2PT/zq8SvYbDUTENEljgelkv+52WkSsqEJ9ZmZtVouEQkRMAial7jeAIQ1MNxoY3RI1mZnZyvybCGZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeVKCwVJW0h6UNIMSdMknZGGd5V0n6QX0t8uhXlGSZop6XlJB5ZVm5mZ1a/MPYXlwLcjoh8wEDhN0g7AmcDEiOgLTEz9pHEjgB2Bg4DLJLUrsT4zM6ujtFCIiDkR8WTqXgzMAHoCw4Dr0mTXAcNT9zDg5oh4NyJeAmYCe5RVn5mZraxFzilI6g3sCkwGNouIOZAFB7BpmqwnMKswW00aVndZp0qaImnKggULSq3bzKytKT0UJHUGxgPfiIhFjU1az7BYaUDEmIgYEBEDunfv3lxlmpkZJYeCpHXJAuHGiLg1DZ4nqUca3wOYn4bXAFsUZu8FzC6zPjMz+7Ayrz4ScDUwIyIuKoyaAIxM3SOB2wvDR0haX1IfoC/wWFn1mZnZytqXuOy9gROAZyQ9nYadBZwPjJV0MvAqcDRAREyTNBaYTnbl0mkRsaLE+szMrI7SQiEiHqL+8wQAQxqYZzQwuqyazMyscf6PZjMzyzkUzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHIOBTMzy7W6UJB0kKTnJc2UdGa16zEza0taVShIagf8GhgK7AB8XtIO1a3KzKztaFWhAOwBzIyIFyPiP8DNwLAq12Rm1ma0r3YBdfQEZhX6a4A9ixNIOhU4NfUukfT8mqzwli/v1dCoTYDX12TZayG3uZndQoOvr2rxc/wRIqmhUU21eauGRrS2UKivhfGhnogxwJjSC5GmRMSAstfTmrjNH31trb3gNq+q1nb4qAbYotDfC5hdpVrMzNqc1hYKjwN9JfWRtB4wAphQ5ZrMzNqMVnX4KCKWS/oacA/QDvhtREyrUjmlH6Jqhdzmj7621l5wm1eJIqLpqczMrE1obYePzMysihwKZmaWcygkkrpKuk/SC+lvlwam21jSOEnPSZohaVBL19pcKm1zmradpKck3dmSNTanStoraQtJD6bndpqkM6pR65pq6nYxyvxfGj9V0m7VqLM5VdDm41Jbp0p6RNLO1aizOVV6WyBJu0taIemoppbpUPjAmcDEiOgLTEz99bkE+HNEbA/sDMxoofrKUGmbAc5g7W4rVNbe5cC3I6IfMBA4bW271UqFt4sZCvRNj1OBy1u0yGZWYZtfAj4dEf2Bn7KWn4Cu9LZAabpfkF3A0ySHwgeGAdel7uuA4XUnkPQxYF/gaoCI+E9ELGyh+srQZJsBJPUCDgGuapmyStNkeyNiTkQ8mboXkwVhz5YqsJlUcruYYcD1kfk7sLGkHi1daDNqss0R8UhEvJl6/072f1Brs0pvC/R1YDwwv5KFOhQ+sFlEzIHsgwHYtJ5ptgYWANekQylXSdqgJYtsZpW0GeBi4HvA+y1UV1kqbS8AknoDuwKTyy+tWdV3u5i6wVbJNGuTVW3PycDdpVZUvibbLKkn8DngikoX2qr+T6Fsku4HPl7PqLMrXER7YDfg6xExWdIlZIcgfthMJTa7NW2zpEOB+RHxhKTBzVhaKZrhOa5dTmeyb1ffiIhFzVFbC2rydjEVTrM2qbg9kvYjC4V9Sq2ofJW0+WLg+xGxopH7JH1ImwqFiNi/oXGS5knqERFz0m50fbtaNUBNRNR+cxxH48fhq64Z2rw3cLikg4EOwMck3RARx5dU8hpphvYiaV2yQLgxIm4tqdQyVXK7mI/aLWUqao+k/mSHQYdGxBstVFtZKmnzAODmFAibAAdLWh4Rf2xooT589IEJwMjUPRK4ve4EETEXmCVpuzRoCDC9ZcorRSVtHhURvSKiN9ltRx5orYFQgSbbq+zdczUwIyIuasHamlMlt4uZAHwxXYU0EHir9tDaWqrJNkvaErgVOCEi/lmFGptbk22OiD4R0Tu9f8cBX20sEGpn8iP7r+5uZFekvJD+dk3DNwfuKky3CzAFmAr8EehS7drLbnNh+sHAndWuu8z2kh1SiPT8Pp0eB1e79tVo68HAP4F/AWenYV8BvpK6RXblyr+AZ4AB1a65Bdp8FfBm4XmdUu2ay25znWmvBY5qapm+zYWZmeV8+MjMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws9/8BTsHskFShKc4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rf_base_best_estimator = rf_base_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "rf_cv_scores = model_selection.cross_val_score(\n",
    "                    rf_base_best_estimator,\n",
    "                    X_train_null,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "rf_train_pred_naive = rf_base_best_estimator.predict(X_train_null)\n",
    "sns.histplot(rf_train_pred_naive).set(title='Random Forest naive model: predicted vs observed')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c49c4be7",
   "metadata": {},
   "source": [
    "#### RF: naive - train / test comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "id": "717e97e3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.068\n",
      "Train MAE:  0.051\n",
      "Test MAE:  0.051\n",
      "Train R-squared: -0.000\n",
      "Test R-squared: -0.000\n",
      "Train RMSE:  0.102\n",
      "Test RMSE:  0.083\n"
     ]
    }
   ],
   "source": [
    "rf_test_pred_naive = rf_base_best_estimator.predict(X_test_null)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(rf_cv_scores): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=rf_train_pred_naive): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=rf_test_pred_naive): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=rf_train_pred_naive): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=rf_test_pred_naive): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=rf_train_pred_naive, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=rf_test_pred_naive, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2ea18fbf",
   "metadata": {},
   "source": [
    "### Random Forest: with features"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6ddcd474",
   "metadata": {},
   "source": [
    "#### RF: with features - fit model - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "id": "01a80c9a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 180 candidates, totalling 900 fits\n",
      "Elapsed time: 80.22 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "rf_regions_mod = rf_baseline_grid.fit(X_train_regions, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c46003ac",
   "metadata": {},
   "source": [
    "#### RF: with features - CV score distribution - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "id": "02d377da",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.70 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f94820dcf70>"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rf_regions_best_estimator = rf_regions_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "rf_cv_scores_regions = model_selection.cross_val_score(\n",
    "                    rf_regions_best_estimator,\n",
    "                    X_train_regions,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "rf_regions_train_pred = rf_regions_best_estimator.predict(X_train_regions)\n",
    "sns.histplot(rf_regions_train_pred).set(title='Predicted vs observed - training data - regions')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f77e565",
   "metadata": {},
   "source": [
    "#### RF: with features - metrics - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "id": "e3c5d793",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.066\n",
      "Train MAE:  0.052\n",
      "Test MAE:  0.053\n",
      "Train R-squared:  0.028\n",
      "Test R-squared:  0.028\n",
      "Train RMSE:  0.100\n",
      "Test RMSE:  0.082\n"
     ]
    }
   ],
   "source": [
    "rf_regions_test_pred = rf_regions_best_estimator.predict(X_test_regions)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(rf_cv_scores_regions): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=rf_regions_train_pred): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=rf_regions_test_pred): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=rf_regions_train_pred): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=rf_regions_test_pred): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=rf_regions_train_pred, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=rf_regions_test_pred, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bb1a4d90",
   "metadata": {},
   "source": [
    "#### RF: with features - plotting observed vs predicted - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "id": "b96ccc86",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))\n",
    "\n",
    "ax1.scatter(y_test, rf_regions_test_pred, c=\"orange\")\n",
    "\n",
    "p1 = max(max(rf_regions_test_pred), max(y_test))\n",
    "p2 = min(min(rf_regions_test_pred), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Random Forest regression on regions - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(rf_regions_test_pred), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title='Random Forest regression on regions - natural scale')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dbf4bc0e",
   "metadata": {},
   "source": [
    "#### RF: with features - performance - predictor: regions\n",
    "\n",
    "The metric that was optimized for, mean absolute error, is worse for random forest regression using region data than it is for naive data. The R-squared value is a little better, and the RMSE is roughly the same."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0926c2c2",
   "metadata": {},
   "source": [
    "#### RF: with features - fit model - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "0b0da697",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 180 candidates, totalling 900 fits\n",
      "Elapsed time: 82.38 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "rf_subregions_mod = rf_baseline_grid.fit(X_train_subregions, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "99c12042",
   "metadata": {},
   "source": [
    "#### RF: with features - CV score distribution - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "id": "fe011347",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.67 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f94812acb20>"
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rf_subregions_best_estimator = rf_subregions_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "rf_cv_scores_subregions = model_selection.cross_val_score(\n",
    "                    rf_subregions_best_estimator,\n",
    "                    X_train_subregions,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "rf_subregions_train_pred = rf_subregions_best_estimator.predict(X_train_subregions)\n",
    "sns.histplot(rf_subregions_train_pred).set(title='Predicted vs observed - training data - subregions')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d0fb720f",
   "metadata": {},
   "source": [
    "#### RF: with features - metrics - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "id": "c5d2905a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.066\n",
      "Train MAE:  0.052\n",
      "Test MAE:  0.055\n",
      "Train R-squared:  0.047\n",
      "Test R-squared:  0.074\n",
      "Train RMSE:  0.099\n",
      "Test RMSE:  0.080\n"
     ]
    }
   ],
   "source": [
    "rf_subregions_test_pred = rf_subregions_best_estimator.predict(X_test_subregions)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(rf_cv_scores_subregions): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=rf_subregions_train_pred): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=rf_subregions_test_pred): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=rf_subregions_train_pred): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=rf_subregions_test_pred): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=rf_subregions_train_pred, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=rf_subregions_test_pred, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c05cc1fc",
   "metadata": {},
   "source": [
    "#### RF: with features - plotting observed vs predicted - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "id": "081d51d8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))\n",
    "\n",
    "ax1.scatter(y_test, rf_subregions_test_pred, c=\"orange\")\n",
    "\n",
    "p1 = max(max(rf_subregions_test_pred), max(y_test))\n",
    "p2 = min(min(rf_subregions_test_pred), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Random Forest on subregions - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(rf_subregions_test_pred), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title='Random Forest on subregions - natural scale')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bebf8b81",
   "metadata": {},
   "source": [
    "#### RF: with features - performance - predictor: subregions\n",
    "\n",
    "The random forest regressor with subregion data performs similarly to the same regressor with region data: worse than the naive model for the MAE metric, better for R-squared, and roughly the same for RMSE. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8edcacdd",
   "metadata": {},
   "source": [
    "#### RF: with features - fit model - predictor: pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "id": "8e65ec19",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 180 candidates, totalling 900 fits\n",
      "Elapsed time: 85.65 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "rf_all_pests_mod = rf_baseline_grid.fit(X_train_all_pests, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3cf09559",
   "metadata": {},
   "source": [
    "#### RF: with features - CV score distribution - predictor: pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "id": "a94eaeba",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.79 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f949b98fb20>"
      ]
     },
     "execution_count": 151,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rf_all_pests_best_estimator = rf_all_pests_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "rf_cv_scores_all_pests = model_selection.cross_val_score(\n",
    "                    rf_all_pests_best_estimator,\n",
    "                    X_train_all_pests,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "rf_all_pests_train_pred = rf_all_pests_best_estimator.predict(X_train_all_pests)\n",
    "sns.histplot(rf_all_pests_train_pred).set(title='Predicted vs observed - training data - all pests')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "50f720a3",
   "metadata": {},
   "source": [
    "#### RF: with features - metrics - predictor: pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "id": "3ccad04f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.063\n",
      "Train MAE:  0.045\n",
      "Test MAE:  0.047\n",
      "Train R-squared:  0.150\n",
      "Test R-squared:  0.132\n",
      "Train RMSE:  0.094\n",
      "Test RMSE:  0.077\n"
     ]
    }
   ],
   "source": [
    "rf_all_pests_test_pred = rf_all_pests_best_estimator.predict(X_test_all_pests)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(rf_cv_scores_all_pests): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=rf_all_pests_train_pred): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=rf_all_pests_test_pred): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=rf_all_pests_train_pred): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=rf_all_pests_test_pred): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=rf_all_pests_train_pred, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=rf_all_pests_test_pred, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "baff4e90",
   "metadata": {},
   "source": [
    "#### RF: with features - plotting observed vs predicted - predictor: pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "id": "3b897957",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))\n",
    "\n",
    "ax1.scatter(y_test, rf_all_pests_test_pred, c=\"orange\")\n",
    "\n",
    "p1 = max(max(rf_all_pests_test_pred), max(y_test))\n",
    "p2 = min(min(rf_all_pests_test_pred), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Random Forest regression on pests - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(rf_all_pests_test_pred), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title='Random Forest regression on pests - natural scale')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d52400d3",
   "metadata": {},
   "source": [
    "#### RF: with features - performance - predictor: pest data\n",
    "\n",
    "The random forest regressor with all pest data performs better than the naive model by all metrics. The RMSE comparison between the train and test sets shows slight underfitting, while the commparison on MAE shows slight overfitting. Given that the metrics show the opposite in terms of fit, and both are slight, this seems like a promising model. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f5a361fe",
   "metadata": {},
   "source": [
    "#### RF: with features - fit model - predictor: mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "id": "85d08c29",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 180 candidates, totalling 900 fits\n",
      "Elapsed time: 82.64 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "rf_mites_regions = rf_baseline_grid.fit(X_train_mites_regions, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "34fa0b2a",
   "metadata": {},
   "source": [
    "#### RF: with features - CV score distribution - predictor: mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "id": "a04f949d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.55 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f94693e11c0>"
      ]
     },
     "execution_count": 155,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rf_mites_regions_best_estimator = rf_mites_regions.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "rf_cv_scores_mites_regions = model_selection.cross_val_score(\n",
    "                    rf_mites_regions_best_estimator,\n",
    "                    X_train_mites_regions,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "train_pred_rf_mites_regions = rf_mites_regions_best_estimator.predict(X_train_mites_regions)\n",
    "sns.histplot(train_pred_rf_mites_regions).set(title='Predicted vs observed - training data - mites and regions')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "78c4ef48",
   "metadata": {},
   "source": [
    "#### RF: with features - metrics - predictor: mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "id": "f72fed93",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.066\n",
      "Train MAE:  0.050\n",
      "Test MAE:  0.051\n",
      "Train R-squared:  0.065\n",
      "Test R-squared:  0.013\n",
      "Train RMSE:  0.098\n",
      "Test RMSE:  0.083\n"
     ]
    }
   ],
   "source": [
    "test_pred_rf_mites_regions = rf_mites_regions_best_estimator.predict(X_test_mites_regions)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(rf_cv_scores_mites_regions): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=train_pred_rf_mites_regions): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=test_pred_rf_mites_regions): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=train_pred_rf_mites_regions): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=test_pred_rf_mites_regions): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=train_pred_rf_mites_regions, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=test_pred_rf_mites_regions, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "390084c3",
   "metadata": {},
   "source": [
    "#### RF: with features - plotting observed vs predicted - predictor: mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "id": "2cd2fc1c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1QAAAEWCAYAAABhbtYTAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAABmBklEQVR4nO3debxc8/3H8dcnG0LssW+xVCwlJY29pUJRSnfElmpJ0EYXpdUFbUqrWlpiqb1i+9mXqH0vIiERISrEEkFiq10k+fz++J5x507mzPfMneXMvff9fDzu4858z/Y9y8yZz/lu5u6IiIiIiIhI9XrknQEREREREZHOSgGViIiIiIhIBymgEhERERER6SAFVCIiIiIiIh2kgEpERERERKSDFFCJiIiIiIh0kAKqOjCz48zskrzzIR1jZlPNbPu881EvZradmT2Tdz6qYWZrmZmbWa8mbGsNM3vfzHo2eltZmdlBZvZA3vnIwsxeMLOheeeju9N9p3PTfSd/3f2+0+qafV80s3vM7AcdXb7LBlTJTf+j5AJ+zcwuNLMl8s5XLcxsezNbkOxT4e/GJm6/aV8+zeTuG7n7PXnno17c/X53X7/wXj+A23P3l9x9CXefn3depGvRfach29d9pxPQfaeyrnbfSb4XZuadj1bSZQOqxB7uvgQwCPgC8Mt8s1MXs5IPZeFvj2pX0MwnJPW8CVrQ1a9ZiehqP6yky9F9pwzdd6Qz032nvrri8ewWXxLu/hpwK+EGB4CZHWNmz5nZe2b2lJl9o2jaQWb2gJn9xczeNrMZZrZr0fQBZnZvsuztwPLF2zOzryfF+e8kRYgbFE17wcyOMrMnzOwDMzvPzFY0s1uS9d1hZstUu49mtkGyrXeSbX+9aNqFZnammY0zsw+AHcxsFTO72szmJPv346L5h5jZBDN718xeN7O/JpPuS/6/kzyl3KpMPo4zs6vM7BIzexc4yMyWSvbzVTN7xcz+ULi5mllPMzvFzN5I8nFE8dPIZJ9Gm9mDwIfA2mY20MxuN7O3zOwZM/tu0fZ3S87ne8m2fp6kL29mNyXH5y0zu79wkyx+kmZmi5jZqWY2K/k71cwWSaZtb2YzzexnZjY72Z/hFc7JKmZ2Q7K96Wb2w5LjdKWZXZzkdaqZDa6wLjezw8zs2WT+35vZOmb2UHKerjSzPsX5TF7/C1gDuDE5Z79I0rc0s/8kx2OyFVU9sXD9P59sZ4aZDUvJ05Bk++8kx+L0Qh6K8jwiyfPbZnaGmVnRef9Lct6fB76Wtu9F5+hoM3sC+MDMekX2YYCZ3Wdtn6kzLKkeZSVPvGs5T0meXkmmPWNmO1baj6zMbGsze9TM/pf83zrLvpVZT6XrfnUzu8bCd8CbZnZ6kr6Omd2VpL1hZmPNbOmU9fewtu/SN5NjtWw9jkFnp/uO7juRz5/uO7rv5Hrfsfh3znAzezrZzvNmdmiSvjhwC7CKtZVar2LhM/+HouU/uyYqHM/U78RI3he18Hl/MzkXj5rZism0Zc3sAgufpbfN7LokfRkLn8c5SfpNZrZahW18P9n/t83sVjNbs2Km3L1L/gEvAEOT16sBU4DTiqZ/B1iFEFR+D/gAWDmZdhDwKfBDoCcwEpgFWDL9IeCvwCLAl4D3gEuSaZ9L1rUT0Bv4BTAd6FOUr4eBFYFVgdnAY4QnmYsAdwG/S9mn7YGZZdJ7J9v4FdAH+EqSp/WT6RcC/wO2Sfa3LzAR+G0y/9rA88BXi/Zv/+T1EsCWyeu1AAd6VTjuxyXHbq9kW4sB1wFnA4sDKwDjgUOT+UcATyXnaBngjuJtAPcALwEbAb2ApYCXgeHJ+82AN4CNkvlfBbZLXi8DbJa8PhE4KzlWvYHtis7nC7RdKyck52cFoD/wH+D3Rcd/XjJPb2A3ws12mZRjcS8wBliU8KNqDrBj0XH6OFlHzyR/D1c4rg7cACyZHItPgDuTc7dUcgwPLHedFO9f8n5V4M1k2z0I1+qbyf4uDrxL27WzcuHYlsnT5sCWyXlYC3gaOLIkzzcBSxNurnOAXYrO+zRgdWBZ4G4qXFvJPkxK5l+s0j4UXcN/IVzf2yb7VPiMrkX7a6xD5wlYn3AtrlK03nU6+H11EPBA8npZ4G1g/+TY7pO8Xy62b2XWW/a6T/ZlMvC35JwvCmybLLNucjwXSa6J+4BTU75bjyR8XlZL5j8buCzP7/48/9B9R/cd3Xd03+lc951K3zlfA9Yh3DO+TLjuCtd2u/Nd9Jn/Q9p3R+nxzPid+EBK3g8FbiR8r/RMroslk2k3A1cQPou9gS8n6csB30qW6Qf8H3Bd0TrvAX6QvN6L8P22AeFa+zXwn4rHsyMnoTP8JSfufcIXvBO+BJauMP8kYM+ikzi9aFrfZB0rET6g84DFi6ZfStuH5jfAlUXTegCvANsX5WtY0fSrgTOL3v+o+ASX5HF7YAHwTtHfdwlf0q8BPYrmvQw4rugiv7ho2hbASyXr/iVwQfL6PuB4YPmSedYi243tvqL3KxK+hBcrStsHuDt5fRfJTS55P5SFb2wnFE3/HnB/yTbPJvkxQLgJHkrywSqa5wTgemDdlGulcGN7DtitaNpXgReKjv9HxftP+GGyZZl1rg7MB/oVpZ0IXFh0nO4omrYh8FGF4+rANkXvJwJHF70/heRHL/Eb29HAv0rWfytwIOHG9g7hS2extPyk5PFI4NqSPG9b9P5K4Jii8z6iaNrOla6tZB++n3EfCp/RvkXTLqHMja2W80QIPGYTrtne1RyrMvt3EG0B1f7A+JLpDyXzVNy3Muste90DWxFu4Kmf5aJ59wIeT/m8PE3yIyB5vzLhBh1db1f8Q/cd3Xfaz6P7Ttt73Xda875T9jsnZf7rgFHlzneSdiHxgOr7kTxNov13YlpA9X3Cg4dNStJXJnxflX3gUDLvIODtovf30BZQ3QIcXDStByGgXDNtfV29yt9e7t6PcFIHUlRFwswOMLNJSVHhO8DGtK9C8Vrhhbt/mLxcghBJv+3uHxTN+2LR61WK37v7AsLThFWL5nm96PVHZd5XasQ8y92XLvq7Mtnmy8m2ivNUvM2Xi16vSSiqfado/39FuAkBHEx44jktKUbdvUJ+yindVm/g1aJtnU14Ekch7ynLpq1vi5K8DyP86IDwhbwb8KKF6jGF6iEnE5423JYUXR+Tkvd25y95vUrR+zfdfV7R+w8pf75WAd5y9/dK1lV8Tl4rev0hsKhVrldcy3VTbE3gOyXHcFvCU6EPCD8eRhDO2c1mNrDcSszsc0mR+WsWqtn8kZJqSCy8j4U8lp734mOepvQ6KLsPtB37D1OWLdbh8+Tu0wk38+OA2WZ2uZkVXytAu96d3jez9zPsZ+k1WJynavYN0q/71YEXS67lQn5XSPblleS8XsLC57VgTeDaonPwNOGHwoop83cHuu+00X1H950C3XfatNJ9J+07BzPb1cwetlAt8R3CNZ52L8iq3THJ8J2Y5l+EYPbypGrfn82sN+He9pa7v126gJn1NbOzzezF5Nq5D1jayrfvXBM4rShfbxFK6lYtMy/QfdpQ3UuInP8CkNSD/CdwBKEazdLAk4SDFfMqsIyFOqQFaxS9nkU4ESTbMsIJfqXjexA1C1jd2jecXaNkm170+mVgRskNsp+77wbg7s+6+z6Em8+fgKuS/S1eRyWl2/qE8NSxsK0l3X2jZPqrhGoXBatnWN+9JXlfwt1HJnl/1N33TPJ+HeHpFO7+nrv/zN3XBvYAfmrl6x23O3+E4zgr436XrmdZM+tXsq5GXgdpSs/by4SnbMXHcHF3PwnA3W91950IN4lphM9KOWcm09dz9yUJP46yfIYgnPfic71G2owp+1FpH14lHPu+RfOXu66gxvPk7pe6+7aEa8YJn5fSeQq9Oy3hobOCmNJrsDhP1exbpev+ZWCNlB9SJyb7sklyXvcj/by+DOxach4Wdfc8rvOWovsOoPuO7jttdN9p04r3nXYstOG7mvD9tWLyfTWOtmNd7nP5AaGUq2ClMvN8tlwt34nu/qm7H+/uGwJbA7sDBxDO0bJWvt3vzwhVJrdIrp0vFbJSZt6XCaXYxed6MXf/T1qeukVAlTgV2MnMBhGKl51Q5QULDTw3zrISd38RmAAcb2Z9zGxbwhdlwZXA18xsxyRa/hnhiz31JNTBI4QL+Rdm1ttCI8k9gMtT5h8PvGuhceBiFhpqbmxmXwQws/3MrH/y5PGdZJn5hOO1gFB/OhN3fxW4DTjFzJa00Ih9HTP7cjLLlcAoM1s1+QAcHVnlTcDnzGz/ZF97m9kXLTSO7mNmw8xsKXf/lFB/eX6yT7ub2brJD41CernuSy8Dfm1m/c1seUJ9/6rHenH3lwnn/EQLjSc3ITyBHVvtuurgddqfs0uAPczsq8m5X9RC49HVLDRU/3ryQ+YTQvWltG5e+xGO5fvJ08SRVeTpSuDHyTaXAdKe3KZJ3Yeiz+hxyTWxFe0/o5+p5TyZ2fpm9pXkxvMx4WltPbrEHUe4xve10Gj3e4QqHzdVs29JHtOu+/GEHwAnmdniyb5vkyzWj3De3zGzVYGjKuT1LGB0cmMk+dzsWcvOdzGnovtOge47uu/ovkPL3ndK9SG0r5wDzLPQWcXORdNfB5Yzs6WK0iYBu1noFGIlQklaJR3+TjSzHczs8xZKl94lVDWfn3z2bwHGWOiEoreZFQKnfoTj9Y6FzpN+V2ETZwG/NLONku0tZWbfqZSnbhNQufsc4GLgN+7+FKHu70OEi+LzwINVrG5fQn3wtwgn5OKi7TxDeKL7D0Kj1T0I3ejOrcNulJWs++vArsk2xwAHuPu0lPnnJ/kaBMxIljmX0MgUYBdgqoVi4tOAvd39Yw/FwaOBBy0Ug26ZMYsHED6cTxEa119FeAoF4enEbcATwOOEH5PzSPmC8FBEvjOwN+Epz2uEpzOLJLPsD7xgoTh3BOFcAKxHaHj8PuG8j/HyY4D8gfCl+AShQfljSVpH7EOoNz0LuJZQ3/72Dq6rFicSbtbvmNnPky/zPQlP9uYQnsQcRfg+6EH4MTaLcH1/GTgsZb0/J3wW3iOcxyuqyNM/CcX1kwnH+JpqdiiyDxCq42xFaDD8hyRvn6SsrqPnaRHgJMLn5zXC0+lfVbMf5bj7m4SnbT8j5P8XwO7u/kYySzX7Vva6L/oOWJfQ/mMmocoNhHYsmxE6FLiZyufmNEKj9dvM7D1Cw/otqtzlLkv3nXbz676j+47uO21a6r5TKrnmf0wIQt8mHPMbiqZPIzwIeD45x6sQquFNJrSVuo3IuanxO3Elwmf6XUJV83tpewixPyHAmkZob3Zkkn4qoXORNwj3qn9XyNu1hM/45cnn+knCd12qQk8eIi0heQpylruXVnkS6TAzuwKY5u6Vnkh1Sl1530SaQfcdaQR9N3cv3aaESlpTUvVjt6Rq06qEJ6/X5p0v6dyS6jjrJFV9diE8Vbwu52zVRVfeN5Fm0H1HGkHfzd1blxupWDodI1QxuoJQt/VmQv1xkVqsRKjOsRyhOttId3883yzVTVfeN5Fm0H1HGkHfzd2YqvyJiIiIiIh0kKr8iYiIiIiIdFCXrPK3/PLL+1prrZV3NkREurWJEye+4e79885HK9J9SkQkf/W6T3XJgGqttdZiwoQJeWdDRKRbM7MX885Dq9J9SkQkf/W6T6nKn4iIiIiISAcpoBIREREREekgBVQiIiIiIiIdpIBKRERERESkgxRQiYhIt2Zm55vZbDN7MmW6mdnfzWy6mT1hZpsVTdvFzJ5Jph3TvFyLiEiryDWgit2IKt3ERERE6uRCYJcK03cF1kv+DgHOBDCznsAZyfQNgX3MbMOG5lRERFpObgFVxhtR2ZuYiIhIvbj7fcBbFWbZE7jYg4eBpc1sZWAIMN3dn3f3ucDlybwiItKN5FlCleVGlHYTExGRBpkzB0aNgnffzTsnLWNV4OWi9zOTtLT0sszsEDObYGYT5syZ05CMiohI8+UZUGW5EWW+WelGJSJSmwUL4OyzYf31YcwYuO++vHPUMqxMmldIL8vdz3H3we4+uH///nXLnIiI5CvPgCrLjSjzzUo3KhGRjnv8cdh6axgxAjbZBCZPht13zztXLWMmsHrR+9WAWRXSRUSkG8kzoMpyI9LNSkSkgd59F448EgYPhuefh4svhrvvhg3VtUKxG4ADko6StgT+5+6vAo8C65nZADPrA+ydzCsiIt1Irxy3/dmNCHiFcCPat2SeG4AjzOxyYAvabmIiIlIDd7jySvjJT+C11+DQQ+GPf4Rllsk7Z81nZpcB2wPLm9lM4HdAbwB3PwsYB+wGTAc+BIYn0+aZ2RHArUBP4Hx3n9r0HRARkVzlFlCl3YjMbEQyPfUmJiIiHffss3D44XD77bDZZnDddTBkSN65yo+77xOZ7sDhKdPGEe5VIiLSTeVZQlX2RpQEUoXXqTcxERGpzscfw4knwkknwaKLwt//DocdBj175p0zERGRzivXgEpERJrj1ltDqdRzz8E++8App8DKGoRCRESkZnl2SiEiIg32yivwne/ALruEkqjbb4dLL1UwJSIiUi8KqEREuqB58+Bvf4OBA+HGG+GEE+CJJ2Do0LxzJiIi0rWoyp+ISBfz0EMwcmQYS2qXXeD002GddfLOlYiISNekEioRkS7izTfhhz8MA/S+8QZcdRWMG6dgSkREpJEUUImIdHILFsAFF4TqfRdcAD/9KTz9NHzrW2CWd+5ERES6NlX5ExHpxJ58MlTve+CBUDJ15pmwySZ550pERKT7UAmViEgn9P77cNRRMGhQKI067zy4/34FUyIiIs2mEioRkU7EHa67DkaNgpdfhoMPDgP1Lr983jkTERHpnlRCJSLSScyYAXvsAd/8Jiy9dKjmd+65CqZERETypIBKRKTFffIJjB4NG24I99wDp5wCEyfCNtvknTMRERFRlT8RkRZ2111w2GHwzDOh175TT4XVVss7VyIiIlKgEioRkRb02muw336w447w6adhPKmrrlIwJSIi0moUUImItJD58+GMM8KYUv/3f/Cb34Su0XfdNe+ciYiISDmq8ici0iImTIARI0L7qB13DIHV+uvnnSsRERGpRCVUIiI5e+cdOPxwGDIEXnkFLr0Ubr9dwZSIiEhnoIBKRCQn7nDJJSFwOussOOIImDYN9tkHzPLOnYiIiGShKn8iIjmYNi303nf33fDFL4ZOJzbfPO9ciYiISLVUQiUi0kQffgjHHgubbAKPPQZjxsBDDymYEhER6axUQiUi0iQ33QQ/+hG88ALsvz+cfDKsuGLeuRIREZFaqIRKRKTBXnoJvvEN2GMPWGyxUM3v4osVTImIiHQFCqhERBrk009DKdQGG8Ctt8If/wiTJsH22+edMxEREakXVfkTEWmA+++HkSNh6tRQMvX3v8Naa+WdKxEREak3lVCJiNTRnDkwfDh86Uvw3ntw3XVwww0KpkRERLoqBVQiInWwYAH8859hTKlLLoGjj4annoI998w7ZyIiItJICqhEupMZY+G6teDSHuH/jLF556jxYvvc0WNStNykv+3GNoPncMgh8PnPh3ZSJ50Eiy9eZV7vGAqXWtvfHUOz5XP8YXBZr7DMZb3C+0rptey3iIiItKM2VCLdxYyxMP4QmP9heP/hi+E9wIBh+eWrkWL7PGMsPDwc/NO26Q8Pb5seWe977/fgt1edwt9v/THL9XuLi/7yH/b/6daYdSCvdwyF2Xe2T5t9Z0hfZ3j6fsx5EKaf2baMzw/vX78X3ntq4XSA/tt0v2tBRESkQVRCJdJdTD627Qd0wfwPQ3pnUKlEJW1abJ8njGoLpgr805BegU86lv/7z24MPGoap906ih/u8E+mnbw+B6yzb8eCKVg4mCpOr7Qfz51TfrniYKrY9LM7/7UgIiLSQlRCJdmMPyz8cPP5YD1hnUNgyJi8cyXV+PCl6tKbbcbY8IP+w5eg7xqw6ei20pIZY+GR78OCueH9hy+G9wVppS2xff70zfLTC+ll8jR9/jCOOO4sbn1iFwat+TjXHPlNtlh3fLLet2vbzzQfvlhdekULWv9aEBER6URUQiVx4w8LVYV8fnhfqDpU3B5DapelTUst7V76rhFPz6tdTaFq3ocvAt4WFBW2P3FUWzBVsGBuSK9U2pJlnyvl6ZHvf5anj995jeOPeo6NN5rPf57dmtP2/zGP/v6LbcEUhIcNhWXLHccZY+Gh/dvv50P7ZzjOaV/VPSpMq6DPstWli4iISCoFVBKXVqUoLV2qFwsoss5TyaajoWff9mk9+4b0eqw/plIHCbEqaHNTSpLmvlm59Ca2zz1Teo3ouXi7IO62J3bi88dM4birf8teg29g2skD+fEu/6BXz/ntl/P5Sbusg9ofx4cPStIPDmntFwrp/TYsn5d+GwILyk9jAR2qaFCahVi6iIiIpMoloDKzZc3sdjN7Nvm/TMp855vZbDN7stl5lCI+v7r07qhSsJBlnixtWmpt9zJgGAw5B/quCVj4P+SctupmWdbf0RKsWClnLVXQCqVC5dIHDIMBB7bNYz3D+8I+91y0/LI9F4W5b/LKW6vwvb9fzlf/dBtmzm3H7MTlh3+TVZZ5LS0z8OgI8Hntk31ekv5J+cX8E/j49fLT0tI/MzcyvYxP36ouvRsws13M7Bkzm25mx5SZvoyZXWtmT5jZeDPbuGjaC2Y2xcwmmdmE5uZcRETyllcJ1THAne6+HnBn8r6cC4FdmpUpSVHpB6tkqxIZDSgytJGpR7uXAcNgrxdg3wXhf3Hbndj6aynBipVyxqrmWZ/y061P5YB/xlh47tz2x/25c9vynFLyNe+jdzj1llFscNTTXP/Ynhz/rd/yxImbsNPn7yisvPw2cZj3fvlJaekFsfZc9dRjkerSuzgz6wmcAewKbAjsY2alRYa/Aia5+ybAAcBpJdN3cPdB7j644RkWEZGWkldAtSdwUfL6ImCvcjO5+31A931k2irWOaS69HroTGPkZKkSGZsnS9BaS3ugLGLtamopIYuVcsaq5nlKKUxaekGsF78yx/3hZ7fgi795lJ9ccipbf+4/PHnSxvz2m79n0T4ppUud0YKPq0vv+oYA0939eXefC1xOuE8V25DwABB3nwasZWYrNjebIiLSivIKqFZ091cBkv8r1LpCMzvEzCaY2YQ5c+bUnEEpMmQMrDuyfbWpdUc2rpe/RrflqbcsVSJj82RZxxLrlp8nLb1asXY1WUrIUqs1pvUlnqTHqiN2VKzUp+j4vvX+Mhx63llsffx/mP3uCvzfj7/NLb/YlXVXeq62PLSTVqqr0t6crQq8XPR+ZpJWbDLwTQAzGwKsCayWTHPgNjObaGapT5p0nxIR6Zoa1m26md0BrFRmUkMGOnH3c4BzAAYPHqym1fU2ZEzzukmvVBLSiEFHO9KNdbWsZ/mgqThIrTQdYM495dedll6tWPDRd43yVRMLJWSFao0FxQPJZukFYcCwHAaV7YH7Ai6670COuuxk3v5gGY7c5VSO/9bx9Fvs3QZsL63d4XzC861ynU+kpUsdlYv4Sy/ak4DTzGwSMAV4HCg0ltvG3WeZ2QrA7WY2Lalh0X6Fuk+JiHRJDSuhcveh7r5xmb/rgdfNbGWA5P/sRuVDOqFmjpGTtTSs1iqIsWqT9SjlyqLSfsSqHcaq5dXaG2Qjqnn2Wa5i+tSZA/ny7+9l+DkXst5KzzLxD5vz1/1+1qBgKqZST36VdOC52Ao7Vpfe9c0EVi96vxowq3gGd3/X3Ye7+yBCG6r+wIxk2qzk/2zgWkIVQhER6SbyqvJ3A3Bg8vpA4Pqc8iGtqJ5thWI/0rP2bFcp6Oq7Zkp+i9Jj1SazrKPWzkFi+xEL2GLV8iotH8v7jLHw8PCSrsaHF52vSJXBNGt8t2zyB/334+ijYdCvJjH1lY345w9+wAO/3ZZN13yi8vqyyHIuy+no+d3qwmiWSlYIQ+9YuJv2fhuGdOhcbRjr41FgPTMbYGZ9gL0J96nPmNnSyTSAHwD3ufu7Zra4mfVL5lkc2BlQz7QiIt1IXgHVScBOZvYssFPyHjNbxczGFWYys8uAh4D1zWymmR2cS26luWIlIVllKX3K0rteLOjKmt8hY2CfebCvh//FVSizrKPWzkFi+5ElEKjUS2ClwWf7b19+UiE91nlERwdOevHK9nM7XDdhTzb41s/5859h/23/xbSTB/KDHc6jR4861cDq6PXb0RLIAcNgq0vaB7pbXVJpQ+Ez8OEL7ZM/fCGkd7Y2jHXg7vOAI4BbgaeBK919qpmNMLMRyWwbAFPNbBqhN8DCxbki8ICZTQbGAze7+7+buwciIpKnhrWhqsTd3wQWqluSVJvYrej9Ps3MlzRJrM1S8bhItbRrytIWK0vbpVgVxKz5rbTfA4bBcxfA7Dvb5l9uq/brKARgz53TVuqzziHZ27bF9mPT0eGHc/ExqyqQ7UX5MZF6wduTyi9SSI+13+qzXPkuztPSS5cHXpizJj+66B/c9PgebLzaFC69fzW2fekHlA/KLCU9gwHDYM6D7c9TYeyrh/ZLXy52LVaaVq792eRjU9q8rRkPrpvZhrFFuPs4YFxJ2llFrx8C1iuz3PPApg3PoIiItKxcAirpxgpPvws/2ApPv2HhoKrWH29Z2mJlKRWIdcYA8fzOGAuPfB8WJAHHhy+G94Vlxx/WPpiC8H78Ye0Dplo6B4ntR6VAIJO0Lszn1j7GUgcLqADmzuvNX27+OX+47tf0sAWcvO/PGfXV0+i97adwaQ0rTpM29lX/bSov15ESqkrTKgXID+1ffplK7RQb0YZRRESkC8iryp90V7WMZVStLG2xIp0WALDKbuXnKU6PtTmZOKotmCpYMDekQ/YOHWLbSe22nPBDukfJALk9+rSVQM0YCzMuah8IzLioNap6dTAgu3vq9mz6y8kce+Uf2XXTW3j65A34+ddOoXeveRWXq0m0+mKKStdi75RpaemQBPkHtm+3VwiQK302Gj3emYiISBejgKrRul/j7sqa2YNflrYsWQooZo0rP08hPUubk7RqaYX0LKUTse0Uui0vDoimn9k+qFpQsp3i91k76OjQ9VyhfVUDvP6/Fdh/zMV85Y9388mni3DzUbtx9ZHfZvXlZpZsftGUbKWkZ9HR0ri576Wne8qgwmnpkF5SNmNs5c9GLPAWERGRdhRQFat38NMNG3dHNfPpd5bBYj99q/yyxemxILBZpW6x7cRKuSaOYuFxkOa3lZLF9rOm67mjXYJXZ/6CHoy5fSTr//wZrnj4exy75x958k8bs9ugW0rmTEptFnyckq2U9FotskqF9ApVJue9X35SIb1cyWSlkrJoj40lTxpK34uIiMhnFFAVNCL4aWb1ts6iXj34ZVWxVzqyBXixeZpV6hbrkTBWyhUrJYvtZy3Xc0e7Ei+o1KX4uiMBmDhjM7b63UMcfuEYNh8wkSmX/ZE/7HsifRf5aOHlei0W32YjSq9W37O69CzSSiZjJWVpn43Jx5YPxLrz95aIiEgFCqgKGhH8NLN6W2eRpdSombIEeLF56lHqlingiFSbq3Wcqk1Hg/UuWbZ3237Wcj0vsW516aUqdBn/zufG8KMrr2XIb8bz0ptrMPawfbnjbyey/neOi5fsVNLh0qsKY2ZNP7v8pLT0gkptqKafVX5aR+l7S0REpCoKqAoa8SNCjbvLi5Ua1VOsGmeWAC82Tz1K3bJ0fBGrNhcb66nH4uWnt0svrdpV9L7PsuWXT0svNvuu6tJLlRkY2dcZyaXTxzBwvY8444avc9hOY5h28kD23eYy7K2HcqxaW6lhXgerPi4zqEJ6ldXx0jq/KND3loiISFXUbXpBlq6xq1XzuD5Sk6xdtGdRqVv0LONQxcYYinV8kcX70yun91oU5n6w8PReSRW2iaPAS3q/83khfcAwmJdSMlNIr7SPqd17VxEMFHUZP20aHH443HUXDF73v9x85PfZfMBjbfN2xqq1lY7f7LvLL5OW/tmyvdqfU+sFm59WeRl9b4mIiFRFJVQFjWjb02rV21pFs3o+zNprXT3azsVK3WLtm+pRQhpbR6wNVWz6gjLBWHF6R8ZRqsb4w/jooiX49V5/YJON5zJx/IeccQY8/LvN2gdTBWltzlpVhWqNHe/Uo7T6YVp1xCL63hIREamKSqgKspQyFMwYm22+wnr1Q6RNPUuNYrIEKZWCrnrmp++aKSWgSRupLCWkvZcr39FAoX1NI0pZq1FpH2sNbsYfxrirXuCIC59gxpy12W+bf3Hyvkex0uBvwnNWvqCrYslYFkbKilPSa1QYsLl4YOV1Dgnp08/sQD5J71widm3re0tERCQzlVAVy9K2R12h16bVBvZtVgP8WAlorEMIgMGnlZ9ncFKFK1M7rApig8fGBkGutP2eKe230tKLvPwyfPOInfjayeNYpPcn3PWrHfjXYQew0tKvtwUf5dRcMpZlkLIqxY7xkDGwzzzY18P/QpBVyQpfqS4P6lxCRESkrhRQVUtdodem1Qb2bVYD/CzVqKykOlbp+wHDYMsL2q9jywva1lFrO6yKHR8Ai6xYfnohvdL2h5xN2epnQ5Le7cqMz/TpvF785d/Hs8EGcMukXRj93V8x+cRN2WGje9pmKpTklGM9od+G5aelpWfWJz29Uucfa363/LS09LaF09PT2s6lHRd1LiEiIlJXqvJXLXUpXJtmVkvLUo2zmQ3wK1WjmnwsLCgZ2HXB3IWrZ1VaR63X5px7Kqe/91T56YX06PZLq6YVBVhzX2+3yIPPbM2I88/iyZmf52tfg38M3YgBK8wos3KrXEK1x1S4caP2ee+3YUivSYVBeBd8Wn7Sgg/jQW9qdeIKbajSjrvPD9eyOpcQERFpKJVQVUtdCtem1Qb2bZUG+LFBe7OIXZuxsa5qrTpXafsTRrFwULAgSW/bxhvvLcf3zzmPbU94kP99tBTX/mQvbrwRBqz0avl191gknq89poYqdIW/rMFUrIpjmh5909MrBZ2VqhNXOnepx33Nytd2szqHERER6eIUUFWr2QFBV9MqAUxpnpo1LlaaWgflhfi1WWsbq1q2X64zDfgsfYH34ty7D2b9nz/Dvx7Yn6O+9mee+vOG7PXFm0LNxw4PsluDpQdVl16w4KP09EpBZ6XqxJWObaXzmnZtzxgLDw9vH7w9PFxBlYiISAcooKpWKwYEnU29Apiu9IS9Hh0rDBgGAw5sN/gtAw6soo1VWpfaSXqsPVIHPxuTJ8O2f5zED889l41Wncrjo7/An/c9miUW/SB9sOJqdPQ66ejYT5Wq51UKfiqVXlU6th1pOzdhVPkeAAslhiIiIpKZ2lB1hLoUzl8zu19vhli36lnMGAszLmoLwnx+eN9/m3BMom2cIr3azU8Zh6o4Pe2z0We5hca5eu+jJfjdtX/m7/+GZRZfiQsOOYgDv3RR+744Ch0uVOoyPq30C2q8Tjo49lOlAXpfvLL8Mi9eGW9fmHZsO9J2LlJiKCKSWTVDyUhtmnGs7xgKs+9se7/CjjD0jury0Q2vCZVQSefU1XpbrEdV0tgxqbX9Xy2dXhRVk3OHqx75Fhsc9TR/u3kkBx8Mz5z8OQ768kULdWz42bpjXcanyeM6qTRAb6VApqNVMtWuU0TyoqFk2ovViKilZk0zjnVpMAXh/R1Ds+ejHvnshDWQFFBJNq12cXe13hbrUZU0dkxqDdpq+eGe9BT43Otrs9ufx/Gdv1/F8v3e4KHjt+Hss2HZFfpVXnesy/g0sWPSa4ny09PSs+i/DQt/tfZI0ivoaLf3HTmvHe1wQ0SkWFd7uFmLLIFGLW1Xm3GsS4OpcumxfGTJ5x1D4VJr+6smYGtRCqgkrhUv7q74VL7WtmWxY1Jr0FZDQPbJ3J6ccM1v2OjoqTz43204df9RTPjDYLZc9z/Z192R4xM7Jl88C6yk5rP1CumVxn6qZPKxlO3RMHbT6+hDgo6c181Pgx4lY2n16BPSRUSy6moPN2sRCyRqbbvaKsc6lo9Yr8WxUrBOGqQroJK4Vry41dviwmoNSioNSFtYtgMB2R13wOePmcLvrj6BPTe7nqdP3oBRu/ydXj3n89lXUKxDjUp6p5Sq9F4ufkwGDIMtLywp+bowpK97aPn1rnsokNb7Ys/KN5tKea3lIUG1weaAYbDF+e33e4vzu3wddxGps674cLOjYoFGrW1XW+VYx/IR67U4VgrWKoFjlRRQSVwrXtzqbXFhtR6TLc6mbFW1Lc5uv42MP9xnzYJ99oGddgKnB7cevTNX/HhvVl12VtHqFwv/0zrUyFIKWql9VZZjkrZPQ8bAuiPbB3nrjgzppPW+OL/yzaZSXlttjDYRkZiu9HCz1vZPjQ54Nh1d/v5Rz2O9wo7x9Ng5b+SYli1MAZXEterFrR+EC6vlmAwYBltd3D742Oriqo/rvHnw97/DwIFw7bVw3HEw5cTPs/Mmty8884Kk5DNLKWjazSzWvqqWY9J/G1hstbDexVaLt4OCyjebSnmtJSButTaOItI9dKaHm5W+J2eMhUe+375pwyPfr66jhVigkaXtauy7vLTnpoV6csqg0jaG3rFwUFXay1+sRkmlgeiz6KRBurmndZXceQ0ePNgnTJiQdza6jtKupyFc3K36pSm5eeQRGDkSHn8cdt4ZTj8d1luP8KWd1i38Xi+EL/ay3bZbCITyuAYrbfOhAyjffXoP2Hd+c7uMbeHPp5lNdPfBuWaiRek+FTH+MHjunPBU23qGXjKHjMk7V9JZxb4nr1p+oaE9gBDsfPuN+D2seDtp3/2FoG3B3Lb5e/Rpq24dy2PWPNRyHLKuo9J+jD8Mpp+58HKFGh6X9qJ8LY+esO+88LKJn/963adUQiVxnekJlOTi7bdhxAjYait47TW44gr497+TYAriT5xipaB5tOOruM3IGFXNLD1txTaOIrUo/CArrgI8/cyQLvXXHUq4Y9+T5YKp4vSsTR8qffcPGAZrH9y+ZGftg9vmieUxax4qnc963C8mjmofTEF4PzHpXCPaY22FKvOF/He0CUCOFFBJNqpeJ2W4w0UXwfrrwz//CaNGwbRp8N3vltREiAXlsYArj3Z8lbZZa5WGemrFNo4itXjunOrSpbJYVbdm9OKbd9BW6/dkPZo+xAKFWB57L1t+enF67HzGeuArrKPSuapX8Jmmkz4kVEAlIh0ydSpsvz0cdBCssw5MnAh/+xssuWTKArEnd5UCrjza8VXaZivV8W7VNo4iHVVro3ZpE/uB3Ywfr80I2mrtMKJSD6zQ8YHXi8WOdZ+UgKmQntZcqji91vNZ61hZkC3wq6STPiRUQCUiVfngAzjmGBg0CKZMgbPPhgcfDO9rUingyiOAiXUu0SrVYFspuBOph1i3y5JdvaqRNTIPUFsJVqxDCYh/Tw4+jYUjFkvS6fjA68Vix3r+x+WnF9JjJUOQrQSqklrHyoJsgV8lnfQhoQIqEcns+uthww3hT3+C/faDZ56BQw6BHo3+JskjgIlts1WqwbZScCdSD+scUl16VxYLNMYfBpf1gkst/C9tZxb7Ed+MH6+xPNRaghVr0wPx78k5D7Jwx0iepGfYhyxix3r+B+WnF9IzPWioMZqpdawsyBb4VdIzZUzMtPQW0SvvDIhI63vhBfjxj+HGG2GjjeC++2C77ZqciULX4l19mx3RWfIpkkWhN6/u3stfaW9qhZIXKN+bWqHzDmg7Vn3XSOkZLvkRv+no8r2+1bOEO5aHSiVYWb7Xsv6Ar/Q9Wand3pAxobpauaCitBpbpV7+Nh1dvne8rMc6U1XYtJ67C+k9SO2htlW891R16S2ihY6giLSauXPhpJNCqdSdd8Kf/xy6RG96MCUi3cuQMbDPPNjXw//uFkxBvOQlS+cdsbY/zSjh3nR0CByKFQcSrdBmJhasLEipjlecnqX9UelQRdUMXdRrierSy1n30Eh6WligcCEmlyNkZsua2e1m9mzyf5ky86xuZneb2dNmNtXMqqjAKSK1uuee0C7ql7+Er34Vnn4ajjoKeveOLSkiIjWLlbxkKbHI0vanGdWXKwUStVY7jHUokUmkqlysOh7E2x9NPrb89KwdRsx7v7r0cvpvA5RWHexZNGh9ZEgQSVUxoEoCn9S/GrZ7DHCnu68H3Jm8LzUP+Jm7bwBsCRxuZhvWsE0RyWD2bDjgANhhB/joo1DN79prYY3Wbg8qUhMz28XMnjGz6Wa20D3JzJYxs2vN7AkzG29mG2ddVlpU3l151ypLm5ospT+xdli1igUStXasE+tQIoteKe1z0tLLibU/ip6LJpQOTT6WhceBmt92LjINCVJrrxMRK+xYXXqLiJ2licCE5P8c4L/As8nriTVsd0/gouT1RcBepTO4+6vu/ljy+j3gaWDVGrYpIhXMnw9nnRXGlLr8cvjVr0LX6LvvnnfORBrLzHoCZwC7AhsC+5R5gPcrYJK7bwIcAJxWxbLSapo1/lJNIj9cs3TeESv9acYgyrFAIku1w0rBb6xDiSzqUfoTE+tOfIUdyk9PSy/HFqmcHusFcNPRYCXVUKx3SXAba6cV0Sel5LCQPvSOhYOnFXYM6S2sYkDl7gPcfW3gVmAPd1/e3ZcDdgeuqWG7K7r7q8k2XgVWqDSzma0FfAF4pMI8h5jZBDObMGfOnBqyJtL9PPYYbL01jBwJX/gCTJ4Mo0dD377xZUW6gCHAdHd/3t3nApcTHvwV25BQowJ3nwasZWYrZlxWWk2nGDw08sN1yBhYd2RbiZT1DO+L25vF2lA1YxDlLFX6KlU7jHWLXo99iJb2ZSg9igUKsYKddyaVn56WXk60nVWG/bCSjJa+j4kdhyzx2NA7QvvJwl+LB1OQvRzxi+7+WYVbd78F+HKlBczsDjN7ssxfVTcaM1sCuBo40t3fTZvP3c9x98HuPrh///7VbEKk2/rf/0LvfV/8YujJ75JLQucTG2yQd85EmmpV4OWi9zNZuEbEZOCbAGY2BFgTWC3jstJqmtERQq1VCrNUv4p13hFrQ9WMQZRrHRQ31jlHPfYhto5oZw7A5qeV73xj86TqYaxNXK3djQN8+lYkPdJGavKx5Y918YOGWPAZOw7RPHZOWQOqN8zs12a2lpmtaWbHAhXPsLsPdfeNy/xdD7xuZisDJP9nl1uHmfUmBFNj3b2WEjERKeIeqvUNHAinnw4jRoQxpYYNq/5hlEgXUO6qL32OehKwjJlNAn4EPE5o65tl2bAR1aRoHY0ef6keVQrrMWh3LHBsxiDKtQ6KGws06rEPseB1yBjoV1KTt9+G7QPYAcNgi/PbV13c4vy20rZmHOtar+ssAwPHqprGjkMnHbg3JmtAtQ/QH7g2+eufpHXUDcCByesDgetLZzAzA84Dnnb3v9awLREp8t//ws47wz77wKqrwiOPwBlnwNJL550zkdzMBFYver8aMKt4Bnd/192Hu/sgQhuq/sCMLMsWrUM1KVpFPYKVSupRpbAeXZrHfrzWYxDlWElco0sDs+5DpXzGStHGH7bwOEjvPbVwW7NKVRdjpWD1GNC21us6S9AX7SmQyseh1hLLFpUpoHL3t9x9FLCdu2/m7ke6ey1lcycBO5nZs8BOyXvMbBUzKzyy2AbYH/iKmU1K/jr30RbJ0UcfwW9/C5//PIwfH0qmHnkkVPcT6eYeBdYzswFm1gfYm/Dg7zNmtnQyDeAHwH1JNfTostKCGj3+UtYgIhaM1NqleewHdpZ2WJVkKYlrdLfoWfYhls9YKVrWdlqVzmcsWKmxrwcgfl1H2zdlqD4Z6ykwptYSyxbVK8tMZrY1cC6wBLCGmW0KHOruHeoGxt3fBBbq/9DdZwG7Ja8foG59MIp0b7fcAkccAc8/D/vuC6ecAiutlHeuRFqDu88zsyMIHTD1BM5396lmNiKZfhawAXCxmc0HngIOrrRsHvshVRowrDFjLkEIFspVnyoOIgo/8gslWYUf+YW81UNhPZOPDcFc3zVCMFW8/iFjOj5wcqWSuMI2Nh3dfj+h+m7RHx7evut1692+W/T+24Qf5B++BIut1r60JEs+YwFwlkAjdj6jgwenjHWVlp6m0nW9+Wnw0P60j9KsrX1T3zVTrtuiKpG1ljhmWX7G2MrXbNZ5mihrlb+/AV8laTfl7pOBLzUqUyJSHzNnwre/DbvtFgbkvfNOGDtWwZRIKXcf5+6fc/d13H10knZWEkzh7g+5+3ruPtDdv+nub1daVrqBWqqQQfN6GmzkwL1ZfhzXWho4YBis84P2JVDr/KBt+SylZLF8xkrRslSFi53PWOlQM8S6mM9y3dZa4hhbPsv5bMFhDzKPFubuL5ck1bELGBGpp08/DaVQAwfCzTfDH/4QukL/ylfyzpmISBdQaxUyaE5Pg42W9cd1LUHdjLHw3Lntx8p67ty2Y50lMI2NAbXp6PI90xVK0bK004p16BCt0teEgX1jVRezXLe1toGKVUPNcj5bcNiDrGfp5aTan5tZHzP7OWGgXRFpMQ8+CJtvDj//OXz5y2Fw3mOPhUVSxvsTEZEqxX7QZQmWukJvZ43u3ANgwqj21f0gvJ+QdJue5Vgv+Lj8PMXpXhLxFL9/97/ll09LL6fWLs3rIVbtMMuxrLUNVKzEMkseWvBhRNaAagRwOGFsjZnAIKCOw2iLSK3eeAN+8APYdlt4+2245hq46SZYe+28cyYi0sXUWoUMmhOMQO3jYVXS6M49AD5N6Ta9kB4rfQKYn9IOqZA++djyQVshQJ59Z/nl09Jj+SmXXq9u1Ws531mu23oEM5VKLLPkoQUfRmQNqNZ392HuvqK7r+Du+xEa6IpIzhYsgPPOC9X7LrwwlEw9/TR84xsaU0pEuqlGBhEQ/0GXJViqRzAS288ZY0OHDsVVEx8eXv+gqlFttLJIu89Vc/9rRolHLJ/1GKC41vOd5bptdDCTJQ/NehhRhawB1T8ypolIEz3xBGy3XSiZGjgQHn8cTj4Zllgi75yJiOSkGQ3WYz/oBgyDAQe270hhwIELBxu1ti165Pvt9/OR77ffz1h1uWapJcCNdeYwN6UqXVp6Oc0o8Yjlsx4lVLWe7yzXbaPHkcryoKEZJaNVqthtupltBWwN9DeznxZNWpKFR/USkSZ57z047jg47bQwIO/558OBB0KPOrZdFRHplLJ05V2rWHfkM8bCjIvad6Qw46LQnXdxHmJdP1eaPnEULJjbPl8L5ob0wjyx6nLNUGv38Gt8F6afWT4doM+yMLfM/vQprmLXg/JtkZKb5hLrlu9UYol14/nLKtaVfqyEqvdy5c9b8ThdtZ7vtA5Aiq/bZowjlWVIg0YOe9ABsZ9ffQhjT/UC+hX9vQt8u7FZE5FS7nD11bDBBvDXv8Lw4fDMM+G/gikREZrXYL1S6VKWXshiJWmx6eWCiErpabJUG6xleq09sr10ZeX0TAPiRjp8mHNP+clp6eXEBiDedDRYSTmG9Wor1YyVUA0+jYV/tvdoPx5XzLojK6dnKeGK9WbYTVX8Cebu97r78cCW7n580d9f3f3ZJuVRRAiD8n7ta2FcqeWWC735/fOf4bWISCaNblvUClqhwXqWoC4WaNSja+hYdbksQV2lNjlZqh1mHcg17bqMBY7R3vOIH4dY6VCWMaQGnxYGHC5WPADxnAfB55Wsf17bGFCZ2lCVNsQqed9z8fLrKKQPGROCp+IqfeuObBvYOUsJVz2qJo4/DC7rBZda+D++A/3ctdh3WdZn2uea2dKFN2a2jJnd2pgsiUixTz6B3/8eNtoI7r8/lExNnAhbb513zkSkU2nBwTAbohUarNejt7R6lLQVqsWlpceCtliJRaVqhwX1GMi1kizHel5Kt+mF9FiQsPlp5cep2ryodCg2APH0s8tvo5Ded82U/UjSJ4xi4SFg57cvPeq5aPl1FKcPGQP7zIN9PfwvBFNZ1dp5xvjDQhXO4mqF08+sLqhqwe+yrAHV8u7+TuFNMkL8Cg3JkYh85s47YZNN4Le/hT32gGnT4Cc/gV4VWz+KiJTRgoNhNkQrNFivR29p9Shpi7V3iQVtsRKLLNUO6zGQayVZjvWClG7TC+mxgXsHDIMtzm9/TW1x/sLt3cq1m/vsR36k2mFscOEspUdZOuiotXQoFvjFxAYXzqIFv8uyBlQLzOyzT7CZrUl6rVURqdGrr8K++8LQoTB/PtxyC1x5Jay6at45E5Fc1VLNpQUHw2yYvLvyzhLUxQKBaKCQ9hOuKL0e42XVqtaBXGPV7eoRQMeqwhW2U+maqseP/AXzK7+PiZ3PWOlQrB0YZAtgKwVt9egevgW/y7IGVMcCD5jZv8zsX8B9wC8bly2R7mn+fPjHP0IX6FdfHUqmpkyBXXbJO2cikrtmVI2S+on9AI8FArHpPRcrv93i9Ng5j3WBHQtmsvwAh9oGco1VW4TQDumjmYCH/4V2SaX5LVWcXmtVuFhnDT1S2jcV0iemVOkrrj4ZEzufsdKhZQaVn16cHrsuY0FbPdpgteB3WaaAyt3/DWwGXAFcCWzu7mpDJVJHjz4KQ4bAj38MW2wBTz4Jxx8Pi6XcM0Wkm2lG1ShprixBV9r00muhXHrsnMeqBMbaDsU6Ysii1jxmaZOTJSirWaTEcIuUNlSF9Hr02vhiSo+IhfRY6VDW3g4rXZfTzyq/jkJ6rHplFi34XVYxoDKzgcn/zYA1gFnAK8AaSZqI1Ojtt+Gww0IQ9eqrcPnlcOutsN56eedMRFpKrdVcWqFtkdRPlqf0sYFaY6UqA4bB2ge3X37tg9uXolXqiKGgUlXVaB4j132WNjn1GDspWt020kYKygenWfVaIp4ebWcVCfqyVser2A4r0o99luqVMS34XRZr2v4z4IfAKWWmOfCVuudIpJtwh0sugZ//HN54I5RMnXACLLlk3jkTkZYUGxg0ixYbDFNqsMpu5Qe8La72FR1gODLgbWz5LAMYF7pWL/QGWOhaHbKto9YBcQvbLCfr2Em1Dk4MoSS5XI+IWQecXpASqKSll9ULmJuSTghuyh3P4up4hRLBgkKJIGQPioaMqb5KZakW+y6LjUP1w+T/DmX+FEyJdNBTT8EOO8ABB8CAATBhApx6qoIpEamgBau5dGt5j4OTpdQlWk00UqpSj7GyYl2rx9YRaxeURa3tdurR4US0hDlSehTrqRDAUkq8PksvF0wVpWepjlePXvq6oIolVGb2zUrT3f2a+mZHpGv78MMwptRf/gJLLAFnnQU//CH0yNo9jIh0X4WnsZOPDT/C+q4RgqkWekrbbdSjxKJWWaqA1lpNtB5jZcXaBsVKj+pRXa/WnuXq0atcj77lg6IehYckGaoMxpSOGRZLL1UoNXrunHBsrGcIpopLk+rRS18XFKvyt0fyfwVga+Cu5P0OwD2AAiqRjG68EX70I3jxRTjwQPjzn2EFjeYmItVosWou3ValEoviqm6NDH6zVAGNzdN7ufLtbgq99MWWjwYJGcSqmbVCF9n1qG674KPK6bFzgVG+fZIVvY60X4pV8YT6VMeLafRnIwexKn/D3X044Uxs6O7fcvdvARs1JXciXcCLL8Kee8LXvw6LLw733gsXXqhgSkSkYRpdHS/2I7/WLu6zyFIFNDZPrJe+2PKxIAHiXavHSjya1UV2pWumHtUOYyVQljL5s/RYsFSHPNRD7Hw347ORg6wVjdZy91eL3r8OfK4B+RHpMubOhT/9CTbcEO64A046CR5/HL70pbxzJiLShTXjB1vsR3492tzEZOnpLMtYV1te0H76lhdkHwsryw/0WNDWd83yqyikx4K6rGNhVTJjLDw8vP018/DwtmumHtUOY+a+VV16ObGxrmLHOotYe7TY+W7GZyMHWQOqe8zsVjM7yMwOBG4G7m5gvkQ6tXvvhS98AY45BnbaKXRCcfTR0KeKHlJFRKqWd0cJraAZP9hiP/KbVU0tNo5VlnmyrCNVpCOFwvorBW2xYxnrVj3LWFgr7Fg+m4X0CaMWbmfkn4Z0yHY+1x1Zfp609FJ9lq2cniVw3OJsFj4nPdrGuqpHxzaxEsVYV/qtUIWzAbIO7HsEcBawKTAIOMfdf9TAfIl0SrNnh/ZR228PH3wAN9wA110Ha1bx8EdEpEO6aFWaqjXjB1us5KZZ1dSyqCXIjl1TPVJGnk9LL2fAMOi7Vvu0vmu1b4tWrlv1Qh5iARvA0DsWDqpW2DGkQ3z8pliwA/HxlWKlQ7EafWumDEJcnD5gGKx7aEkeDq2ixDGDLFX6Kp2vVvps1FE1fYs9Btzs7j8BbjWzfg3Kk0ins2ABnH02DBwIl14aSqamToU99ogvKyJSF120Kk3VmvWDrVLJTqt0cV9rkB27phZ8uPAypemx6nR3DIX3nmq//HtPhfQseYBspWxLfq59oLFkFS1XsjZfGjIG9pkH+3r4X9y5Q6wdViyoy1LtMBbM1EOsrVfsfLXKZ6POMgVUZvZD4CogKTNkVeC6BuVJpFN5/HHYaisYMQI22QQmT4YTTwwdUIiINE0XrUpTtVb4wVaPkoB6qDXIjl1TWYLXWHW62XeWX0chPct1HSuFKwxGWxxoTD8zpAP0SSl1KaR/mtKOKS29nOcvrpwea5uUZXDi2PmuRyl2rK1X7Hy1ymejzrKWUB0ObAO8C+DuzxK6Uhfptt59F0aNgsGDYcYMuPhiuPvu0AmFiEjTddGqNFVrlR9sNbVNqpNag+zYNZUleI2VvNSahyxBQmww2s1PA0oDmp5JOtA7pcpfWno5sYF5Y22TsgxOHDvf9SjFjp2PLN9DrfDZqLOsAdUn7v7Z8Mpm1ovq+mkU6TLc4YorQvW+f/wDDj0UnnkG9t8fLK0oXESk0VqhZKZVdMEfbB1Sa5CdpcOIRgevsTxkCRKyDEbboyRgKX4f7dK8DqJtrDLsQ+x816O0L3Y+uun3UNaA6l4z+xWwmJntBPwfcGPjsiXSmp59Fr76Vdh7b1h5ZXj4YRgzBpZZJu+ciUi31yolM9I6Nh0NPUq6l+3RJ/uP26xds1cKXmPV6WI98MXykCVIiJXuTD4WFsxtP23B3LagbG5KaVpaekfEApGeKe0IitNj7bTqUdqXpSv+bvg91CvjfEcDPwCmAIcC44BzG5UpkVbz8cehXdRJJ8Gii4aSqZEjoWfKd7SISC4GDOvyP1ykSu6V38fErqkZY0Pg8eFL4Yf5pqPbz7/5afDI99sHLD36tFWnG3pH6ICiuC1VcQ98sTz0XrZ89cHi6njrHBLaTJVa55DwP9o+qQflx9yqom+3HouXr/ZXGCNqwDB47oL2x2G5rdr2e37KIMrF6S9eWX6eF68MHWRsOjoESMUlellL+0qDaH3PtBMNqMysB/CEu28M/LPxWRJpLbfeCocfDs89B/vsA6ecEkqnREREWtrkY8t3CFH6A7mjCiUahR/hhRINaF9iURooLL/dwt2ad1SW6niF3vaeOydUkbOeIZgqpFvP8lXqPivZyjCAcczyW5bvgGP5LcP/8YctPH32nSF9yJhseYi1VxswDOY82P44FI/pVY+ObbJcE1nWUSlIb0HR0NrdFwCTzaybtWqV7u6VV+A734FddgklUbffHrpEVzAlIiKdQqN7fszSfqlSoFBQy1hZsV7nCvpvA4utBlj433+btmlZ2ifVas49ldOnn11+elp6RzRjjKhaO76IdbPforKWVa4MTDWzO83shsJfRzdqZsua2e1m9mzyf6EWKGa2qJmNN7PJZjbVzI7v6PZEqjFvHvztb6HTiZtugt//Hp54AoYOzTtnIiLSNLX8yG8Vje75MUvAFuthr9auvLPsY2wbsQ4hCtXySpWmV7pmokFbHUrBYu3VmjFGVK1BfKyb/RaVNaA6HtgdOAE4peivo44B7nT39YA7k/elPgG+4u6bAoOAXcxsyxq2KRL1n//A5pvDT38K220XBuf99a9hkUXyzpmIiDRNPcbraQX1+IFcKUjIEszEAolaSzSy7GOtgUSvRctvuzi90ddMLOiD0C6tXCckhfZqzRgjqtYgvtZu9nNSMaBKSomOBL4DDAQedPd7C381bHdP4KLk9UXAXqUzePB+8rZ38qeu2qUh3nwTfvhD2Gab8Pqqq+Dmm2HttfPOmYiINF09xutpBbX+QI4FCa1QopFlH2sNJLJUK6z1mon14hfrwQ9Cfrc4v/1+bHF+235kGU+r1iEH1G16WRcBgwm9++1KbaVSxVZ091cBkv9lBwk2s55mNgmYDdzu7o+krdDMDjGzCWY2Yc6cOXXKpnR1CxbA+efD+uvDBRfAz34GTz8N3/qWxpQS6U7MbBcze8bMppvZQrUmzGwpM7uxqBr68KJpL5jZFDObZGYTmptzaYhGtz1qplp+IMeChFYo0YDQ0cJHMwEP/+c8WP02Kh2nLIFI7JqJlTAtl1IJq5D+/MXlp5emV9oP/6T8OorTa63qWus1Eau22KJivfxt6O6fBzCz84DxWVdsZncAK5WZlPnxjrvPBwaZ2dLAtWa2sbs/mTLvOcA5AIMHD1ZJlkRNmRK6Pn/wQdh6azjzTNhkk7xzJSLNZmY9gTOAnYCZwKNmdoO7P1U02+HAU+6+h5n1B54xs7FFg97v4O5vNDfn0jB91yjflXa92h51FlkCy1q70I515R0z/rD2XaL7/Lb3hV78at1GlkAkds3E8hDrtKJcl+uV0suZ937l9Hr00FeYt6PXRKyb/RYVK6H6rFWYu8+rZsXuPtTdNy7zdz3wupmtDJD8nx1Z1zvAPcAu1eRBpJz334ejjoIvfAGmTYPzzoP771cwJdKNDQGmu/vzSYB0OaFqejEH+pmZAUsAbwFV3RelE+mm1ZYW0uhOLaD2Eo1Ypxf12EYsEIH4NRPLQ716GqylhKkVqrrGqi22qFgJ1aZm9m7y2oDFkvdGaOa0ZAe3ewNwIHBS8v/60hmSJ4Cfuvs7ZrYYMBT4Uwe3J4I7XHstjBoFM2fCwQeHgXqXXz7vnIlIzlYFXi56PxPYomSe0wn3rllAP+B7ybAiEIKt28zMgbOTGhMLMbNDgEMA1lijm5V0dDaFH2+dbCycuqu1ZAfCD+KyJTdFVeBqKdHIGog0ejDaLNdMpTxEx8LKMLhwrISp93IpgyAn1elapaprJxw4uGIJlbv3dPclk79+7t6r6HVHgykIgdROZvYsoYrFSQBmtoqZjUvmWRm428yeAB4ltKG6qYZtSjf2/POw++6hbdQyy8ADD8C55yqYEhGg/NCgpVXHvwpMAlYh9Dx7upkV7oPbuPtmhLbGh5vZl8ptxN3PcffB7j64f//+dcl4l9UKXZbX2ji/K6hHG6lGl/Z9FnBkTO+I3intd0rTa7lm1jmkcvq6h5afXpweK2EafBpY7/bTrXdIh+aUSHZRWbtNryt3f9Pdd3T39ZL/byXps9x9t+T1E+7+BXffJKkqeEIeeZXO7ZNPYPRo2GgjuPdeOOUUmDgx9OYnIpKYCaxe9H41QklUseHANUkPtNOBGYTeb3H3Wcn/2cC1hCqE0lFdpcvyrqLWwLIeQVklsUCkHmKBSD0MGQPrjmwLBK1neF9oB9Y/5YdLcXqW3gy3vKD9udjygrZzoaquHRar8ifSad11Fxx2GDzzTCiZOvVUWG21vHMlIi3oUWA9MxsAvALsDexbMs9LwI7A/Wa2IrA+8LyZLQ70cPf3ktc7E8ZslI6q9JS9O5YSdQWNrMLVfxuYfjbtq8P1WDgAmTG241U4m1UFdMiYtgCqVNrAthNGteUjS2cqlc6Fqrp2mAIq6XJeey10f37ppWEcqXHjYNdd886ViLQqd59nZkcAtwI9gfPdfaqZjUimnwX8HrjQzKYQqgge7e5vmNnahF5oIdxTL3X3f+eyI11Fq7TjkM5h8rEs3LZoQfsAvB691+XdrifLgLf1aPOW9352UgqopMuYPx/OOguOPRY++gh+8xv45S9hscXyzpmItDp3HweMK0k7q+j1LELpU+lyzwObNjyD3Ym6LJdqZAnAu0upp0qYcqOASrqECRNgxIjQPmrHHWHMGPjc5/LOlYiIVK0eT9ml+8gSgGcJumqpEtgMfZaDuWVKqUoHvFUJUy5y6ZRCpF7eeQcOPxyGDIFXXoHLLoPbb1cwJSLSaTW6EwPpWrJ0pBDrva5VOkKp1LvlGt8tv0xaep5aoZfOJlMJlXRK7jB2bGgr9cYbcMQR8Pvfw1JL5Z0zERGpmZ6yS1ZZqrnFSj1boUpgrJ3XrHHll0tLz0s92qt1QgqopNOZNi303nf33aFk6pZbYLPN8s6ViIiI5CIWgMeCrlboCCUW1LVCHrNoheA0BwqopNP48MMwptTJJ8Pii8OZZ8IPfwg96zh2n4iIiHRBlYKuVugIJRYwtUIes+gsgV+dqQ2VdAo33RQG5/3jH2HvvUMp1YgRCqZERESkRq0woG2snVcr5DGL2H50UQqopKW99BJ84xuwxx6h+/N77oGLL4YVV8w7ZyIiItIltEJHKLGAqRXymEVnCfzqTFX+pCV9+in87W9w/PGhA4oTT4Sf/hT69Mk7ZyIiItLl5N0RSpbONfLOYxbddCwsBVTScu6/H0aOhKlTQ8nU3/8Oa62Vd65EREREGqgzBExZdJX9qIKq/EnLmDMHDjoIvvQleO89uP56uOEGBVMiIiIi0roUUEnuFiyAc86B9dcPY0sdfTQ89RR8/et550xEREREpDJV+ZNcTZoUqvc9/HAomRozJvTmJyIiIiLSGaiESnLx7rtw5JGw+ebw3HNw0UWhBz8FUyIiIiLSmaiESprKHf7v/+AnP4FXX4VDDgljSy27bN45ExERERGpnkqopGmmT4dddoHvfQ9WWAEeegjOOkvBlIiIiIh0XgqopOE+/hiOOw423jgEUaedBo8+CltskXfORERERERqoyp/0lC33QaHHx5Kp/beG045BVZZJe9ciYiIiIjUh0qopCFeeSVU7fvqV8EsBFaXXaZgSkRERES6FgVUUlfz5sGpp8IGG4SBeU84AZ54AnbaKe+ciYiIiIjUn6r8Sd08/HAYU2rSpND5xOmnwzrr5J0rEREREZHGUQmV1Oytt+DQQ2HrrWHOnNAt+rhxCqZEREREpOtTQCUd5g4XXgjrrw/nnRfGlnr6afj2t0O7KRERERGRrk4BlXTIk0/Cl78Mw4fDeuvBxImhB79+/fLOmYiIiIhI8yigkqq8/z784hfwhS/A1Knwz3/CAw/AppvmnTMRERERkeZTpxSSiTtcdx2MGgUvvwzf/z786U+w/PJ550xEREREJD8qoZKoGTNgjz3gm9+EpZaC++8PbaYUTImIiIhId6eASlLNnQt//CNstBHccw+cfDI89hhsu23eORMRERERaQ2q8idl3X03HHYYTJsWSqZOPRVWXz3vXImIiIiItJZcSqjMbFkzu93Mnk3+L1Nh3p5m9riZ3dTMPHZXr78O++0HX/kKfPIJ3HwzXH21gikRERERkXLyqvJ3DHCnu68H3Jm8TzMKeLopuerG5s+HMWPCmFJXXgnHHhu6Rt9tt7xzJiIiIiLSuvIKqPYELkpeXwTsVW4mM1sN+BpwbnOy1T1NmABbbgmHHw6bbw5TpsAf/gB9++adMxERERGR1pZXQLWiu78KkPxfIWW+U4FfAAtiKzSzQ8xsgplNmDNnTt0y2pW98w4ccQQMGRK6Qh87Fu64I5RSiYiIiIhIXMMCKjO7w8yeLPO3Z8bldwdmu/vELPO7+znuPtjdB/fv37+mvHd17nDppTBwYKjmd/jhofOJffcFs7xzJyLSfGa2i5k9Y2bTzWyhauhmtpSZ3Whmk81sqpkNz7qsiIh0bQ3r5c/dh6ZNM7PXzWxld3/VzFYGZpeZbRvg62a2G7AosKSZXeLu+zUoy93CtGkhgLrrLhg8OHQ6sfnmeedKRCQ/ZtYTOAPYCZgJPGpmN7j7U0WzHQ485e57mFl/4BkzGwvMz7CsiIh0YXlV+bsBODB5fSBwfekM7v5Ld1/N3dcC9gbuUjDVcR99BL/+NWyyCUycGEqmHn5YwZSICDAEmO7uz7v7XOByQlvfYg70MzMDlgDeAuZlXFZERLqwvAKqk4CdzOxZwlO9kwDMbBUzG5dTnrqscePC4LyjR8P3vgfPPAMjR0LPnnnnTESkJawKvFz0fmaSVux0YANgFjAFGOXuCzIuC6itr4hIV5VLQOXub7r7ju6+XvL/rSR9lrsv1FG3u9/j7rs3P6ed28svh0F5v/Y1WGSRUM3vX/+CFVfMO2ciIi2lXOtRL3n/VWASsAowCDjdzJbMuGxIVFtfEZEuKa8SKmmgTz+Fv/wFNtgA/v1v+OMfYfJk2GGHvHMmItKSZgLFw5evRiiJKjYcuMaD6cAMYGDGZUVEpAtTQNXFPPAAbLYZHHVUCKCmToVf/hL69Mk7ZyIiLetRYD0zG2BmfQjtdm8omeclYEcAM1sRWB94PuOyIiLShTWslz9prjlz4Oij4YILYPXV4dprYc891Q26iEiMu88zsyOAW4GewPnuPtXMRiTTzwJ+D1xoZlMI1fyOdvc3AMotm8d+iIhIPhRQdXILFsB558Exx8C778IvfgG//S0svnjeORMR6TzcfRwwriTtrKLXs4Cdsy4rIiLdhwKqTmzy5NBb30MPwXbbha7QN94471yJiIiIiHQfakPVCb33Hvz0p2EMqWefhQsvhHvvVTAlIiIiItJsKqHqRNzhqqvgyCNh1iw45BA48URYdtm8cyYiIiIi0j2phKqTmD4ddt0Vvvtd6N8/VPM7+2wFUyIiIiIieVJA1eI+/hhOOCFU5/vPf+DUU2HCBNhyy7xzJiIiIiIiqvLXwm6/HQ4/PLST+u534a9/hVVXzTtXIiIiIiJSoBKqFjRrFuy9N+y8c2g3deutcMUVCqZERERERFqNAqoWMm8enHYaDBwI110Hxx0HU6aEwEpERERERFqPqvy1iEceCWNKPf54CKBOPx3WWy/vXImIiIiISCUqocrZ22/DiBGw1Vbw+utw5ZXw738rmBIRERER6QwUUOXEHS66CNZfH/75Txg1Cp5+Gr7zHTDLO3ciIiIiIpKFqvzlYOpUOOwwuO++0P35bbfBoEF550pERERERKqlEqom+uADOOaYEDxNmQLnnAMPPqhgSkRERESks1IJVZNcfz38+Mfw0ktw0EHw5z9D//5550pERERERGqhEqoGe+EF+PrXYa+9oF+/UM3vggsUTImIiIiIdAUKqBpk7lw48UTYcEO4885QIvX447DddnnnTERERERE6kVV/hrgnntCpxNPPw3f+AaceiqssUbeuRIRERERkXpTCVUdvf467L8/7LADfPQR3HgjXHONgikRERERka5KAVUdzJ8PZ54JAwfCFVfAr34Vukbfffe8cyYiIiIiIo2kKn81euwxGDECHn00lEydcQZssEHeuRIRERERkWZQCVUH/e9/8KMfwRe/CC++CJdcEjqfUDAlIiIiItJ9qISqSu5w+eXw05+GNlMjR8Lo0bD00nnnTEREREREmk0BVRWeeQYOPzyURG2+OdxwQyihEhERERGR7klV/jL46CP4zW9gk01CW6nTT4dHHlEwJSIiIiLS3amEKuKWW+CII+D552HffeGUU2CllfLOlYiIiIiItAKVUKWYORO+/W3YbTfo3TtU8xs7VsGUiIiIiIi0UUBV4tNPQynUwIFw882hw4nJk+ErX8k7ZyIiIiIi0mpyqfJnZssCVwBrAS8A33X3t8vM9wLwHjAfmOfugxuZr1degV13hSlT4Gtfg3/8AwYMaOQWRURERESkM8urhOoY4E53Xw+4M3mfZgd3H9ToYApCdb511oFrroEbb1QwJSIiIiIileXVKcWewPbJ64uAe4Cjc8rLZ3r2hGuvzTsXIiIiIiLSWeRVQrWiu78KkPxfIWU+B24zs4lmdkilFZrZIWY2wcwmzJkzp87ZFRERERERWVjDSqjM7A6gXJ94x1axmm3cfZaZrQDcbmbT3P2+cjO6+znAOQCDBw/2qjMsIiIiIiJSpYYFVO4+NG2amb1uZiu7+6tmtjIwO2Uds5L/s83sWmAIUDagEhERERERaba8qvzdAByYvD4QuL50BjNb3Mz6FV4DOwNPNi2HIiLSbZjZLmb2jJlNN7OFOkoys6PMbFLy96SZzU96rMXMXjCzKcm0Cc3PvYiI5CmvgOokYCczexbYKXmPma1iZuOSeVYEHjCzycB44GZ3/3cuuRURkS7LzHoCZwC7AhsC+5jZhsXzuPvJSY+zg4BfAve6+1tFszStR1oREWktufTy5+5vAjuWSZ8F7Ja8fh7YtMlZExGR7mcIMD2572BmlxN6o30qZf59gMualDcREWlxeZVQiYiItIpVgZeL3s9M0hZiZn2BXYCri5Iz9Uir3mhFRLomBVQiItLdWZm0tN5i9wAeLKnut427b0aoMni4mX2p3ILufo67D3b3wf37968txyIi0jLyGti3oSZOnPiGmb1YxSLLA280Kj+dhI5BoOOgYwA6BgW1Hoc165WRBpsJrF70fjVgVsq8e1NS3a8jPdJmuE91hmtQeayfzpBP5bF+OkM+O0MeoUXuU+auIZvMbEJ3b0isYxDoOOgYgI5BQXc5DmbWC/gvoW3vK8CjwL7uPrVkvqWAGcDq7v5BkrY40MPd30te3w6cUGsnSp3h2CuP9dMZ8qk81k9nyGdnyCO0Tj67ZAmViIhIVu4+z8yOAG4FegLnu/tUMxuRTD8rmfUbwG2FYCqxInCtmUG4p16qHmlFRLoXBVQiItLtufs4YFxJ2lkl7y8ELixJU4+0IiLdnDqlCM7JOwMtQMcg0HHQMQAdgwIdh/x0hmOvPNZPZ8in8lg/nSGfnSGP0CL5VBsqERERERGRDlIJlYiIiIiISAcpoBIREREREemgbhlQmdmyZna7mT2b/F8mZb4XzGyKmU0yswnNzmcjZT0Gybw9zexxM7upmXlshizHwcwWNbPxZjbZzKaa2fF55LVRMh6D1c3sbjN7OjkGo/LIa6NU8Z1wvpnNNrMnm53HRjGzXczsGTObbmbHlJluZvb3ZPoTZrZZHvnsKmLHO5ln++S+M9XM7q1m2RbIY9Pumxmu3aOSfEwysyfNbL6ZLZtl2RbJYysdy6XM7Mai++DwrMu2SB6bciwz5HEZM7s2+S4db2YbZ122hfLZrGNZ8X5b6d7UzGP5GXfvdn/An4FjktfHAH9Kme8FYPm885vnMUim/xS4FLgp73zncRwAA5ZIXvcGHgG2zDvvTT4GKwObJa/7Ecbs2TDvvDfzGCTTvgRsBjyZd57rtN89geeAtYE+wOTS8wrsBtySfA62BB7JO9+d9S/j8V4aeApYI3m/QtZl885j8rop981qjwewB3BXqx3LtDy22rEEflX4XgT6A28l87bMsUzLY7OOZcY8ngz8Lnk9ELizmddkrfls8nVZ8X5Lyr2pmcey+K9bllABewIXJa8vAvbKLyu5yXQMzGw14GvAuc3JVtNFj4MH7ydveyd/Xak3lyzH4FV3fyx5/R7wNLBqszLYBJk+D+5+H+Em3VUMAaa7+/PuPhe4nHAsiu0JXJx8Dh4GljazlZud0S4iy/HeF7jG3V8CcPfZVSybdx6bqdrjsQ9wWQeXzSOPzZQlnw70MzMDliB8D87LuGzeeWyWLHncELgTwN2nAWuZ2YoZl22FfDZNhvtt2r2pmcfyM901oFrR3V+F8EMRWCFlPgduM7OJZnZI03LXHFmPwanAL4AFTcpXs2U6DhaqPU4CZgO3u/sjzctiw2W9FgAws7WALxBK6rqKqo5BF7Iq8HLR+5ksHChnmUeyyXIsPwcsY2b3JPeeA6pYNu88QvPum5mPh5n1BXYBrq522RzzCK11LE8HNgBmAVOAUe6+IOOyeecRmnMss+RxMvBNADMbAqwJrJZx2VbIJ7TOb+O0/cjlntVlB/Y1szuAlcpMOraK1Wzj7rPMbAXgdjOblkTMnUKtx8DMdgdmu/tEM9u+jllrqnpcC+4+HxhkZksD15rZxu7eadrR1OnzgJktQbjhH+nu79Yjb81Sr2PQxViZtNLS1yzzSDZZjmUvYHNgR2Ax4CEzezjjsvXQ4Ty6+39p3n2zmuOxB/CguxeedrfSsSwozSO01rH8KjAJ+AqwTpKf+zMuWw8dzmNyr2rGscySx5OA05IHtFOAxwmlaM38nq0ln9A6v43T9iOXe1aXDajcfWjaNDN73cxWdvdXk+LBstUV3H1W8n+2mV1LKEbsNAFVHY7BNsDXzWw3YFFgSTO7xN33a1CWG6Ie10LRut4xs3sITxI7TUBVj2NgZr0JwdRYd7+mQVltmHpeB13ITGD1overEZ7uVjuPZJP1eL/h7h8AH5jZfcCmGZfNO4//beJ9s5rjsTftq9K10rEsKM1jM3+DZMnncOAkd3dgupnNILStaaVjmZbH8U06ltE8JsHdcAidKgAzkr++sWVbJJ+t9Ns4bT/6pKQ3VHet8ncDcGDy+kDg+tIZzGxxM+tXeA3sTCf6AZ1B9Bi4+y/dfTV3X4vwZX9XZwumMshyLfRPSqYws8WAocC0ZmWwCbIcAwPOA5529782MW/NEj0GXdSjwHpmNsDM+hA+5zeUzHMDcEDSo9KWwP8K1SOlalmO9/XAdmbWK6kGtgWhzWKWZXPNY5Pvm5mOh5ktBXyZ9p/pVjqWZfPYgsfyJUKJJBba0qwPPJ9x2Vzz2MRjGc2jmS2dTAP4AXBfErw06zjWlM8W+22cdm9q5rFs4w3u9aIV/4DlCI3tnk3+L5ukrwKMS16vTahDOhmYChybd76bfQxK5t+ertnLX5ZrYRNCcfcThC+O3+ad7xyOwbaEIvMnCFUqJgG75Z33Zh6D5P1lwKvAp4SnYwfnnfc67PtuhF4bnyt8zwEjgBHJawPOSKZPAQbnnefO/Bc73sn7owi96D1JqF6bumwr5bHZ982M+TwIuDzLsq2Ux1Y7lsl34W3Jd8CTwH6tdizT8tjMY5khj1sR7jPTgGuAZZp9HGvJZ5OP5UL3WzLem5p5LAt/lmxYREREREREqtRdq/yJiIiIiIjUTAGViIiIiIhIBymgEhERERER6SAFVCIiIiIiIh2kgEpERERERKSDFFBJrsxsOTOblPy9ZmavFL3vE19D3fOzvZnd1OBtvN/I9TeKmQ1KBnmux7oafpxFRPKg+1rnUc/7mnRvvfLOgHRv7v4mMAjAzI4D3nf3vxSmm1kvd5+XT+6kxCBgMDAu53yIiLQs3dc6lUHoviZ1oBIqaTlmdqGZ/dXM7gb+ZGbHmdnPi6Y/aWZrJa/3M7PxyZO/s82sZ5n1fdHM/mNmk5N5+5nZomZ2gZlNMbPHzWyHMssta2bXmdkTZvawmW2SpB9nZueb2T1m9ryZ/ThJ/72ZjSpafnRhWsp+mpmdnOzPFDP7XpK+spndl+zTk2a2nZn1TI5LYd6flFnfimZ2bbKfk81s6yT9p8lyT5rZkUnaWmb2ZNGyP09u/CT79afkWP032X4f4ATge0m+vmdmXy566vp4YfT0alU4zv3N7HYzeyw5ty+a2fId2YaISJ50X+te9zXpflRCJa3qc8BQd59f+EIsZWYbAN8DtnH3T81sDDAMuLhonj7AFcD33P1RM1sS+AgYBeDunzezgcBtZva5kk0cDzzu7nuZ2VeS9Q5Kpg0EdgD6Ac+Y2ZnAeYQRxU8zsx7A3sCQCvv4zWR9mwLLA4+a2X3AvsCt7j46uZH2TeZb1d03TvZr6TLr+ztwr7t/I1luCTPbHBgObEEYVfwRM7sXeLtCvgB6ufsQC1UhfufuQ83st4SRyI9I8nAjcLi7P2hmSwAfR9aZJu04/w64y91PNLNdgEM6uH4RkVag+1r3ua9JN6OASlrV/7n7/Mg8OwKbE76wARYDZpfMsz7wqrs/CuDu7wKY2bbAP5K0aWb2IuFmV2xb4FvJPHdZqBe/VDLtZnf/BPjEzGYDK7r7C2b2ppl9AViRcNN6s0L+twUuS/bz9eSG8EXgUeB8M+sNXOfuk8zseWBtM/sHcDNwW5n1fQU4IMnvfOB/yX5e6+4fJPt9DbAdcEOFfEG4gQJMBNZKmedB4K9mNha4xt1nRtaZJu04bwt8I0n/t5nFbpYiIq1M97Xuc1+TbkZV/qRVfVD0eh7tr9VFk/8GXOTug5K/9d39uJL1GOBl1m8Z8lBunsK6PilKm0/bw4lzgYMIT8/O78D6cff7gC8BrwD/MrMD3P1twhO/e4DDk+1kkbaface0oLB/xftWms+TgB8QbvgPJ09E2zZsdnhR1YlVqsyjV8i7iEhnpPtaJ7+viaRRQCWdwQvAZgBmthkwIEm/E/i2ma2QTFvWzNYsWXYasIqZfTGZp5+Z9QLuI1SjIKkSsQbwTMmyxfNsD7xReBJYwbXALoQncrdG5r2PUHe7p5n1J9xsxif7MNvd/0mobrGZhbZDPdz9auA3heNR4k5gZJLfnkk1kPuAvcysr5ktTijxuR94HVgheTq5CLB7JK8A7xGqgpBsYx13n+LufwImEKqLfMbdzyj6UTArchzKHecHgO8m6TsDy2TIo4hIZ/ACuq91uvuaSBpV+ZPO4GrgADObRKg28F8Ad3/KzH5NqCfeA/iU8JTrxcKC7j7XQqPYf5jZYoR65kOBMcBZZjaF8FTrIHf/JKliUXAccIGZPQF8CBwYy2iyvbuBdzJU7bgW2AqYTHhC+At3f83MDgSOMrNPgfcJ1R1WTfJSeAjyyzLrGwWcY2YHE57AjXT3h8zsQmB8Ms+57v44gJmdADwCzCDcoGPuBo5JzsOJwLYWGj3PB54CbsmwjnKOo/xxPh64LDl/9wKvEm5+IiKdne5rXfu+Jt2MuZcrNRaRjkhuDI8B33H3Z/POT2eWPGGc7+7zzGwr4Ex3H5RztkREuhXd10TiVEIlUidmtiFwE6GxrG46tVsDuDK5mc8FfphzfkREuhXd10SyUQmViIiIiIhIB6lTChERERERkQ5SQCUiIiIiItJBCqhEREREREQ6SAGViIiIiIhIBymgEhERERER6aD/BxfanCsHPExUAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1008x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 4))\n",
    "\n",
    "ax1.scatter(y_test, test_pred_rf_mites_regions, c=\"orange\")\n",
    "\n",
    "p1 = max(max(test_pred_rf_mites_regions), max(y_test))\n",
    "p2 = min(min(test_pred_rf_mites_regions), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Random Forest regression on mites and regions - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(test_pred_rf_mites_regions), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title=\"Random Forest regression on mites and regions - natural scale\")\n",
    "ax2.axis('equal')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4af2ea87",
   "metadata": {},
   "source": [
    "#### RF: with features - performance - predictor: mites and regions\n",
    "\n",
    "The random forest regressor with varroa mite and region data combined performs similarly to the same regressor with region data: worse than the naive model for the MAE metric, better for R-squared, and roughly the same for RMSE. This indicates that the addition of the varroa mite data to region data doesn't improve the model as I would have anticipated. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "21fdf3fb",
   "metadata": {},
   "source": [
    "#### RF: with features - fit model - predictor: Patsy matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "id": "b1174649",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 180 candidates, totalling 900 fits\n",
      "Elapsed time: 77.26 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "rf_matrix_train = rf_baseline_grid.fit(train_model_matrix, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b9772dea",
   "metadata": {},
   "source": [
    "#### RF: with features - CV score distribution - predictor: Patsy matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "id": "ee010291",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.69 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f949b412220>"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rf_matrix_best_estimator = rf_matrix_train.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "rf_cv_scores_matrix = model_selection.cross_val_score(\n",
    "                    rf_matrix_best_estimator,\n",
    "                    train_model_matrix,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "rf_train_pred_matrix = rf_matrix_best_estimator.predict(train_model_matrix)\n",
    "sns.histplot(rf_train_pred_matrix).set(title='Predicted vs observed - training data - Patsy matrix')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d88c63b0",
   "metadata": {},
   "source": [
    "#### RF: with features - metrics - predictor: Patsy matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "id": "44131292",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.066\n",
      "Train MAE:  0.050\n",
      "Test MAE:  0.051\n",
      "Train R-squared:  0.066\n",
      "Test R-squared:  0.038\n",
      "Train RMSE:  0.098\n",
      "Test RMSE:  0.081\n"
     ]
    }
   ],
   "source": [
    "rf_test_pred_matrix = rf_matrix_best_estimator.predict(test_model_matrix)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(rf_cv_scores_matrix): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=rf_train_pred_matrix): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=rf_test_pred_matrix): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=rf_train_pred_matrix): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=rf_test_pred_matrix): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=rf_train_pred_matrix, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=rf_test_pred_matrix, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "42ae6a4e",
   "metadata": {},
   "source": [
    "#### RF: with features - plotting observed vs predicted - predictor: Patsy matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "id": "b4365027",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 4))\n",
    "\n",
    "ax1.scatter(y_test, rf_test_pred_matrix, c=\"orange\")\n",
    "\n",
    "p1 = max(max(rf_test_pred_matrix), max(y_test))\n",
    "p2 = min(min(rf_test_pred_matrix), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Random Forest on Patsy matrix - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(rf_test_pred_matrix), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title=\"Random Forest on Patsy matrix - natural scale\")\n",
    "ax2.axis('equal')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "179aa168",
   "metadata": {},
   "source": [
    "#### RF: with features - performance - predictor: Patsy matrix\n",
    "\n",
    "The random forest regressor with a Patsy matrix of varroa mite, pesticide, and region data combined only performs better than the naive model on the R-squared metric. The other two metrics are approximately the same as the baseline model. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "09c14003",
   "metadata": {},
   "source": [
    "### Random Forest regression - overall conclusions\n",
    "\n",
    "The one model that stands out random forest regression is the one built with pests as the predictive fields. In comparison to the models built with linear ridge regression, it falls in the middle of the two stand outs there, in terms of metrics, performing slightly better than the Patsy matrix and slightly worse than the linear ridge pest model. The training and test data sets align about as well as the pest data for linear ridge. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af440b1a",
   "metadata": {},
   "source": [
    "## Support Vector regression"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2ba42d24",
   "metadata": {},
   "source": [
    "#### SV - create model object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "id": "d5b1dcbd",
   "metadata": {},
   "outputs": [],
   "source": [
    "sv_base_model = svm.SVR(kernel=\"poly\", degree=3, epsilon=0.1, coef0=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f4cb680",
   "metadata": {},
   "source": [
    "#### SV - create params object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "id": "2c27fd68",
   "metadata": {},
   "outputs": [],
   "source": [
    "sv_params = {'model__C': [0.1, 1, 10, 100, 1000], \n",
    "             'model__gamma': [1, 0.1, 0.01, 0.001, 0.0001, 'auto']}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4aeac34e",
   "metadata": {},
   "source": [
    "#### SV - create pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "id": "5ff37c12",
   "metadata": {},
   "outputs": [],
   "source": [
    "sv_pipeline = pipeline.Pipeline(steps=[(\"model\", sv_base_model)])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d1dca11",
   "metadata": {},
   "source": [
    "#### SV - create baseline grid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "id": "f80e5100",
   "metadata": {},
   "outputs": [],
   "source": [
    "sv_baseline_grid = model_selection.GridSearchCV(\n",
    "                        estimator=sv_pipeline,\n",
    "                        param_grid=sv_params,\n",
    "                        refit=True,\n",
    "                        scoring='neg_mean_absolute_error',\n",
    "                        cv=cross_validator,\n",
    "                        verbose=1\n",
    "                    )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d8019d13",
   "metadata": {},
   "source": [
    "### Support Vector: naive model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3c4fa02c",
   "metadata": {},
   "source": [
    "#### SV: naive - fit model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "id": "6a4ca2d1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n",
      "Elapsed time: 0.90 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "sv_base_mod = sv_baseline_grid.fit(X_train_null, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "acd9f306",
   "metadata": {},
   "source": [
    "#### SV: naive - CV score distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "id": "690bfbae",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.03 seconds\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sv_base_best_estimator = sv_base_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "sv_cv_scores = model_selection.cross_val_score(\n",
    "                    sv_base_best_estimator,\n",
    "                    X_train_null,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "sv_naive_train_pred = sv_base_best_estimator.predict(X_train_null)\n",
    "sns.histplot(sv_naive_train_pred, label='CV scores').set(title='Support vector naive model: predicted vs observed')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "93f9d6a5",
   "metadata": {},
   "source": [
    "#### SV: naive - train / test comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "id": "2793652e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.070\n",
      "Train MAE:  0.058\n",
      "Test MAE:  0.057\n",
      "Train R-squared: -0.004\n",
      "Test R-squared: -0.007\n",
      "Train RMSE:  0.102\n",
      "Test RMSE:  0.083\n"
     ]
    }
   ],
   "source": [
    "sv_naive_test_pred = sv_base_best_estimator.predict(X_test_null)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(sv_cv_scores): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=sv_naive_train_pred): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=sv_naive_test_pred): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=sv_naive_train_pred): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=sv_naive_test_pred): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=sv_naive_train_pred, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=sv_naive_test_pred, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0cfa7448",
   "metadata": {},
   "source": [
    "### Support Vector: with features"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bc625504",
   "metadata": {},
   "source": [
    "#### SV: with features - fit model - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "id": "321b263e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n",
      "Elapsed time: 1.11 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "sv_regions_mod = sv_baseline_grid.fit(X_train_regions, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "81195917",
   "metadata": {},
   "source": [
    "#### SV: with features - CV score distribution - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "id": "14718ec5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.05 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f93f80374f0>"
      ]
     },
     "execution_count": 170,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sv_regions_best_estimator = sv_regions_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "sv_cv_scores_regions = model_selection.cross_val_score(\n",
    "                    sv_regions_best_estimator,\n",
    "                    X_train_regions,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "train_pred_regions_sv = sv_regions_best_estimator.predict(X_train_regions)\n",
    "sns.histplot(train_pred_regions_sv).set(title='Predicted vs observed - training data - regions')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8d1164fa",
   "metadata": {},
   "source": [
    "#### SV: with features - metrics - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "id": "d80a386e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.070\n",
      "Train MAE:  0.057\n",
      "Test MAE:  0.057\n",
      "Train R-squared:  0.023\n",
      "Test R-squared:  0.012\n",
      "Train RMSE:  0.101\n",
      "Test RMSE:  0.083\n"
     ]
    }
   ],
   "source": [
    "test_pred_sv_regions = sv_regions_best_estimator.predict(X_test_regions)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(sv_cv_scores_regions): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=train_pred_regions_sv): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=test_pred_sv_regions): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=train_pred_regions_sv): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=test_pred_sv_regions): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=train_pred_regions_sv, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=test_pred_sv_regions, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "27be4566",
   "metadata": {},
   "source": [
    "#### SV: with features - plotting observed vs predicted - predictor: regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "id": "3676b535",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))\n",
    "\n",
    "ax1.scatter(y_test, test_pred_sv_regions, c=\"orange\")\n",
    "\n",
    "p1 = max(max(test_pred_sv_regions), max(y_test))\n",
    "p2 = min(min(test_pred_sv_regions), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Support Vector regression on regions - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(test_pred_sv_regions), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title=\"Support Vector regression on regions - natural scale\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3b5b07f7",
   "metadata": {},
   "source": [
    "#### SV: with features - performance - predictor: regions\n",
    "\n",
    "The support vector regression model with region as the predictor only performs well on the R-squared metric, but performs the same as the naive model for MAE and RMSE. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e1c532e",
   "metadata": {},
   "source": [
    "#### SV: with features - fit model - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "id": "e7bdd13e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n",
      "Elapsed time: 1.28 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "sv_subregions_mod = sv_baseline_grid.fit(X_train_subregions, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "052e6294",
   "metadata": {},
   "source": [
    "#### SV: with features - CV score distribution - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "id": "692b64fb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.06 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f949ba717c0>"
      ]
     },
     "execution_count": 174,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEICAYAAACwDehOAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAAAfC0lEQVR4nO3deZxcdZ3u8c8jAUIS1EACZmkMSsAA9kQMQQ1IRIZEBglwWaLihG1ABEXEBXDQuHCHiyy5ymVziICyBVmMDioQBMyAQMAYEkIuKCFpOoQYCYRVAt/54/z6cCiqu6vTtXSnn/frVa8+9Tvb93fqVD11zqmuUkRgZmYG8I5GF2BmZj2HQ8HMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHIOhRqTdLmkH6ThPSQtqdN6Q9J29VhXmXWPSuvv14j1r69GbrO0/or3jzrvS9Ml/bwe66oVSRMltdRxfYskTazX+qrJoQBIWirpZUkvSFop6aeSBlV7PRHxh4jYoYJ6jpA0t9rrt9qpxgtnpftHV6etp+KboL4sInaKiDsbXcf6cCi86dMRMQjYBdgV+PfSCXrbO98NmTK9Zv/tbfX2dZI2anQNjeKdtEREPAX8BtgZ8lMKJ0h6DHgste0nab6kNZLukdTcNr+kD0l6SNJaSdcB/Qvj3nIIK6lJ0o2SVklaLekCSWOAi4GPpiOXNWnaTSWdI2lZOpq5WNJmhWV9XdIKSa2Sjmqvf5KmSppX0naypNlpeF9Jj6T6n5L0tXaW8w5J/y7pSUnPSLpS0rtKJjsq1bNC0imFecdLmifp+dSX8wrjPpK26RpJfy4egku6U9KZkv4beAk4vZO+VGWbdUbSZOB04LD0mP25nXrfJ+lISYvT9v2rpOMKyyndP5ZK+pqkBZKek3SdpP5dnTaN/0ahr8eog1NlkraVdFeq8TZgSMn46yU9ndZzt6SdUvuxwOeAb6Tt8KvUfqqkv6TlPSLpwG5s6yGSfp32j79L+oNS2Jb2SWWOWiSdLulvaXt9rmTaiyTdIulF4BOShku6Qdnz8wlJXy5Mv5mkKyQ9mx7Pb5R5PPZOw5tKmpG2fWsa3jSNmyipRdIpyp5HKyQdWVhORc/HqoqIPn8DlgJ7p+EmYBHw/XQ/gNuALYDNyI4kngF2AzYCpqX5NwU2AZ4ETgY2Bg4GXgN+kJY1EWhJwxsBfwbOBwaShcfuadwRwNySGmcAs1MdmwO/Av4jjZsMrCQLsoHA1anu7cr0dQCwFhhdaHsAmJqGVwB7pOHBwC7tbLOjgMeB9wGDgBuBn6Vxo9L6r0n1fBBYVdjG9wKfT8ODgI+k4RHAamBfsjcs/5zuD03j7wSWATsB/YB3ddKXqmyzCveh6cDPS9pK690Y+Bfg/YCAPcnCYpfS/aOwX94PDE99WAx8YT2mnQw8neoYAPyso76mx+c8sn3642kb/7ww/qi0PTdN23h+YdzlpP290HZIqusdwGHAi8Cw9dzO/0H2pmnjdNsDUOG5ul25WtL2Wlfo156pjh0K0z4HTEh1DgAeBL5N9rx+H/BXYFKa/izgLrLnyEhgQZnHo21//x7wR2ArYChwD2++vrTV9b3Un33TPjG4K8/Hqr4e1noFveGWHsAXgDVkL+oXApsVdrS9CtNe1PaAFtqWpJ3s40Br206axt1D+VD4KNkLZb8y9RxBIRTIXkBeBN5faPso8EQangmcVRi3fekTpGT5Pwe+nYZHkz3pB6T7y4DjgHd2ss3mAF8s3N+BLAD78WYofKAw/mzgsjR8N/BdYEjJMr9JCpZC2++AaWn4TuB7lfSl2tusgn1oOuVD4XudzHczcFLp/lHYLw8v2YYXr8e0M0lhmO5v115fgW3IXqQGFtquLu1bYdy707Lele5fTkkolJlnPjBlPbfz94BftlN7JaFQ7Ncs4IzCtFcWxu0GLCtZ/mnAT9NwHhDp/jFlHo+2UPgLsG9h3CRgaaGulym8DpC96Wx7o1TR87GaN58+etMBEfHuiHhvRHwxIl4ujFteGH4vcEo6fF2j7PROE9k7oeHAU5EezeTJdtbXBDwZEesqqG0o6Z1LYZ2/Te2k9RZrbG+dba4GPpOGPwvcHBEvpfv/i+zdypPpFMJH21nG8JL1PEkWCFsX2kprGp6GjyZ7EX5U0gOS9kvt7wUOKdm2uwPD2llmR32p2jZT9kmfF9JtUXvTteMt9Ur6lKQ/plMfa8i29ZCyc2aeLgy/RHZk1dVpS/taug2LhgPPRsSLhbZ820jaSNJZ6XTQ82QvftBBHyT9q9483bqG7Ois7PSF7fyCpG3KTPJDsiPUW5Wdfju1g76UKtev4YX7pc/z4SX74um8uX93dZuWPleK611d8jpQfOwqfT5WjUOhMsUX+eXAmSlA2m4DIuIaskO9EZJUmL7cjt22nG1U/uJ1lNz/G9m7iZ0K63xXZBfGSettqmCdbW4FhkgaS/aCenW+4ogHImIK2aHuzWTvpsppJXviFNe5juyUTJvSmlrTOh6LiM+kdfwf4BeSBpJtk5+VbNuBEXFWYTml26a9vlRtm0X2SZ9B6bZTe5N11p7OI98AnANsHRHvBm4hO6qppRVkpzjaNLU3YZp2cHo82hS3zWeBKcDeZKfvRqX2tj68ZTtIei/wE+BEYMvU54W00+fCdh4UEcvKjF8bEadExPuATwNflfTJNPolsjcCbd5TMnu5frUWF18YXk52VFncFzePiH3T+K5s03LPldZ2pn2LLjwfq8ah0HU/Ab4gaTdlBkr6F0mbk52LXQd8WVI/SQcB49tZzv1kO9ZZaRn9JU1I41YCIyVtAhARb6T1ni9pKwBJIyRNStPPAo6QtKOkAcB3OupAelfyC7J3XVuQXTNB0iaSPifpXRHxGvA88Ho7i7kGOFnZRclBwP8Grit5x3OGpAHKLkQeCVyX1nO4pKGpX2vStK+TnQr6tKRJ6R1p/3Qhrvjkq6gv1d5mFVgJjFLHnzDahOx89ipgnaRPAft0c72VmAUcKWlM6uu325swIp4E5gHfTfvD7mQvvm02B14lu9YzgOxxL1pJdv69zUCyF9tVAOki6s7r2xFlH/LYLr3xats/2/bR+cBn074zmeyUbqm2fu0B7Adc386q7geel/RNZReVN5K0s6Rd0/hZwGmSBksaQRZ67bkG+HdJQyUNIdv+nX58uYvPx6pxKHRRRMwD/g24AHiW7FD2iDTuH8BB6f6zZBfVbmxnOa+TPdm2Iztv2JKmB7iD7GL305L+ltq+mdb1x3TYfjvZeXwi4jdkF/zuSNPcUUFXriZ7t3d9yQv554GlaR1fAA5vZ/6ZZBcs7waeAF4BvlQyzV2pnjnAORFxa2qfDCyS9ALwf8kuDL8SEcvJ3oWeTvYishz4Op3vp+31pdrbrCNtLy6rJT1UboKIWAt8mewF5Vmyd92zu7neTqW+/gj4PVlf702jXm1nls+SnVP/O1lYXlkYdyXZ6Y+ngEfILqAWXQbsmE653BwRjwDnpnWuJPvQwX93ozujyR7HF9IyL4w3/x/gJLLn1BqyT0HdXDLv02TbvRW4iuxC/KPlVlJ4fo4l27//Bvwn2dERZNc2WtK428nemLS3PX9AFrQLgIeBh1JbJSp9PlZN21V7M+sjlH3seSGwaYXXtKwTko4ne3NT7uikV/GRglkfIOnAdDpiMNl1nF85ENafpGGSJij7f50dgFOAmxpdVzU4FMz6huPITsn9hey89PGNLafX2wS4hOwj0HeQfUz2woZWVCU+fWRmZjkfKZiZWa5Xf8HbkCFDYtSoUY0uw2yDt2R19i3dO2zZ476Y1dbDgw8++LeIGFpuXK8OhVGjRjFv3rzOJzSzbpl4+UQA7jzizobWYdUhqd3/4PfpIzMzyzkUzMws51AwM7Ncr76mYGYbrtdee42WlhZeeeWVRpfSa/Xv35+RI0ey8cYbVzyPQ8HMeqSWlhY233xzRo0axVu/eNgqERGsXr2alpYWtt1224rn8+kjM+uRXnnlFbbccksHwnqSxJZbbtnlIy2Hgpn1WA6E7lmf7edQMDOznEPBzKwLli5dys47r/fvBNXMxIkTq/LPvA4Fsx5mRNM2SKrLbURTZ7/cavWwbl3P+RZzf/rIrIdpbVnOYZfcU5d1XXfcx+qynt7svPPOY+bMmQAcc8wxHHDAAaxbt45p06bxpz/9ie23354rr7ySAQMGcOqppzJ79mz69evHPvvswznnnMOqVav4whe+wLJl2U9Oz5gxgwkTJjB9+nRaW1tZunQpQ4YM4S9/+QszZ85kp52ynwGfOHEi5557Lh/4wAf40pe+xMMPP8y6deuYPn06U6ZM4eWXX+bII4/kkUceYcyYMbz88stV6a9Dwcx6vK/89ivMf3p+VZc59j1jmTF5RofTPPjgg/z0pz/lvvvuIyLYbbfd2HPPPVmyZAmXXXYZEyZM4KijjuLCCy/kqKOO4qabbuLRRx9FEmvWrAHgpJNO4uSTT2b33Xdn2bJlTJo0icWLF+fLnzt3Lpttthnnn38+s2bN4rvf/S4rVqygtbWVD3/4w5x++unstddezJw5kzVr1jB+/Hj23ntvLrnkEgYMGMCCBQtYsGABu+yyS1W2i08fmZm1Y+7cuRx44IEMHDiQQYMGcdBBB/GHP/yBpqYmJkyYAMDhhx/O3Llzeec730n//v055phjuPHGGxkwYAAAt99+OyeeeCJjx45l//335/nnn2ft2rUA7L///my22WYAHHrooVx/ffZT37NmzeKQQw4B4NZbb+Wss85i7NixTJw4kVdeeYVly5Zx9913c/jh2U82Nzc309zcXJU++0jBzHq8zt7R10p7P0JW+lFPSfTr14/777+fOXPmcO2113LBBRdwxx138MYbb3DvvffmL/5FAwcOzIdHjBjBlltuyYIFC7juuuu45JJL8hpuuOEGdtjh7V9bXouP7PpIwcysHR//+Me5+eabeemll3jxxRe56aab2GOPPVi2bBn33nsvANdccw277747L7zwAs899xz77rsvM2bMYP78+QDss88+XHDBBfky29rLmTp1KmeffTbPPfccH/zgBwGYNGkSP/7xj/OA+tOf/pTXdtVVVwGwcOFCFixYUJU+OxTMzNqxyy67cMQRRzB+/Hh22203jjnmGAYPHsyYMWO44ooraG5u5u9//zvHH388a9euZb/99qO5uZk999yT888/H4Af/ehHzJs3j+bmZnbccUcuvvjidtd38MEHc+2113LooYfmbWeccQavvfYazc3N7LzzzpxxxhkAHH/88bzwwgs0Nzdz9tlnM378+Kr0uVf/RvO4cePCP7JjGxpJdf30USWvAY34kZ3FixczZsyYuq1vQ1VuO0p6MCLGlZveRwpmZpZzKJiZWc6hYGY9Vm8+vd0TrM/2cyiYWY/Uv39/Vq9e7WBYT22/p9C/f/8uzef/UzCzHmnkyJG0tLSwatWqRpfSa7X98lpXOBTMrEfaeOONu/SLYVYdPn1kZmY5h4KZmeUcCmZmlnMomJlZzqFgZma5moWCpCZJv5e0WNIiSSel9umSnpI0P932LcxzmqTHJS2RNKlWtZmZWXm1/EjqOuCUiHhI0ubAg5JuS+POj4hzihNL2hGYCuwEDAdul7R9RLxewxrNzKygZkcKEbEiIh5Kw2uBxcCIDmaZAlwbEa9GxBPA40B1vgvWzMwqUpdrCpJGAR8C7ktNJ0paIGmmpMGpbQSwvDBbC2VCRNKxkuZJmuf/dDQzq66ah4KkQcANwFci4nngIuD9wFhgBXBu26RlZn/bl55ExKURMS4ixg0dOrQ2RZuZ9VE1DQVJG5MFwlURcSNARKyMiNcj4g3gJ7x5iqgFaCrMPhJorWV9Zmb2VrX89JGAy4DFEXFeoX1YYbIDgYVpeDYwVdKmkrYFRgP316o+MzN7u1p++mgC8HngYUnzU9vpwGckjSU7NbQUOA4gIhZJmgU8QvbJpRP8ySMzs/qqWShExFzKXye4pYN5zgTOrFVNZmbWMf9Hs5mZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZrmahYKkJkm/l7RY0iJJJ6X2LSTdJumx9HdwYZ7TJD0uaYmkSbWqzczMyqvlkcI64JSIGAN8BDhB0o7AqcCciBgNzEn3SeOmAjsBk4ELJW1Uw/o2aCOatkFS3W4jmrZpdJfNrAr61WrBEbECWJGG10paDIwApgAT02RXAHcC30zt10bEq8ATkh4HxgP31qrGDVlry3IOu+Seuq3vuuM+Vrd1mVnt1OWagqRRwIeA+4CtU2C0BcdWabIRwPLCbC2prXRZx0qaJ2neqlWralq3mVlfU/NQkDQIuAH4SkQ839GkZdribQ0Rl0bEuIgYN3To0GqVaWZm1DgUJG1MFghXRcSNqXmlpGFp/DDgmdTeAjQVZh8JtNayPjMze6tafvpIwGXA4og4rzBqNjAtDU8DfllonyppU0nbAqOB+2tVn5mZvV3NLjQDE4DPAw9Lmp/aTgfOAmZJOhpYBhwCEBGLJM0CHiH75NIJEfF6DeszM7MStfz00VzKXycA+GQ785wJnFmrmszMrGP+j2YzM8s5FMzMLOdQMDOznEPBzMxyDgUzM8s5FMzMLOdQMDOznEPBzMxyDgUzM8s5FMzMLOdQMDOznEPBzMxyfToU6vk7xv4NYzPrDWr51dk9Xj1/x9i/YWxmvUGfPlIwM7O3ciiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZrmKQkHShErazMysd6v0SOHHFbaZmVkv1uF/NEv6KPAxYKikrxZGvRPYqJaFmZlZ/XX2NRebAIPSdJsX2p8HDq5VUWZm1hgdhkJE3AXcJenyiHiyTjWZmVmDVPqFeJtKuhQYVZwnIvaqRVFmZtYYlYbC9cDFwH8Cr9euHDMza6RKP320LiIuioj7I+LBtltHM0iaKekZSQsLbdMlPSVpfrrtWxh3mqTHJS2RNGk9+2NmZt1QaSj8StIXJQ2TtEXbrZN5Lgcml2k/PyLGptstAJJ2BKYCO6V5LpTkTzeZmdVZpaePpqW/Xy+0BfC+9maIiLsljapw+VOAayPiVeAJSY8D44F7K5zfzMyqoKJQiIhtq7jOEyX9KzAPOCUingVGAH8sTNOS2t5G0rHAsQDbbOOfuDQzq6aKQiG9iL9NRFzZxfVdBHyf7Cjj+8C5wFGAyi2+nXVeClwKMG7cuLLTmJnZ+qn09NGuheH+wCeBh4AuhUJErGwblvQT4NfpbgvQVJh0JNDalWWbmVn3VXr66EvF+5LeBfysqyuTNCwiVqS7BwJtn0yaDVwt6TxgODAauL+ryzczs+6p9Eih1EtkL9ztknQNMBEYIqkF+A4wUdJYslNDS4HjACJikaRZwCPAOuCEiPD/Q5iZ1Vml1xR+xZvn+DcCxgCzOponIj5TpvmyDqY/EzizknrMzKw2Kj1SOKcwvA54MiJaalCPmZk1UEX/vJa+GO9Rsm9KHQz8o5ZFmZlZY1T6y2uHkl34PQQ4FLhPkr8628xsA1Pp6aNvAbtGxDMAkoYCtwO/qFVhZmZWf5V+99E72gIhWd2Fec3MrJeo9Ejht5J+B1yT7h8G3FKbkszMrFE6+43m7YCtI+Lrkg4Cdif7Sop7gavqUJ+ZmdVRZ6eAZgBrASLixoj4akScTHaUMKO2pZmZWb11FgqjImJBaWNEzCP7aU4zM9uAdBYK/TsYt1k1CzEzs8brLBQekPRvpY2SjgY6/DlOM7O+ZETTNkiq221EU21+T6azTx99BbhJ0ud4MwTGAZuQfcupmZkBrS3LOeySe+q2vuuO+1hNltthKKTfP/iYpE8AO6fm/4qIO2pSjZlt0EY0bUNry/K6rW/4yCaeWr6sbuvbEFT6ewq/B35f41rMbAO3obyb3pD5v5LNzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs5xDwczMcg4FMzPLORTMzCznUDAzs1zNQkHSTEnPSFpYaNtC0m2SHkt/BxfGnSbpcUlLJE2qVV1mZta+Wh4pXA5MLmk7FZgTEaOBOek+knYEpgI7pXkulLRRDWszM7MyahYKEXE38PeS5inAFWn4CuCAQvu1EfFqRDwBPA6Mr1VtZmZWXr2vKWwdESsA0t+tUvsIoPgbfS2p7W0kHStpnqR5q1atqmmxZmZ9TU+50KwybVFuwoi4NCLGRcS4oUOH1rgsM7O+pd6hsFLSMID095nU3gI0FaYbCbTWuTYzsz6v3qEwG5iWhqcBvyy0T5W0qaRtgdHA/XWuzcysz+tXqwVLugaYCAyR1AJ8BzgLmCXpaGAZcAhARCySNAt4BFgHnBARr9eqNjMzK69moRARn2ln1Cfbmf5M4Mxa1WNmZp3rKReazcysB3AomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeUcCmZmlnMomJlZzqFgZmY5h4KZmeX6NWKlkpYCa4HXgXURMU7SFsB1wChgKXBoRDzbiPrMzPqqRh4pfCIixkbEuHT/VGBORIwG5qT7ZmZWRz3p9NEU4Io0fAVwQONKMTPrmxoVCgHcKulBScemtq0jYgVA+rtVuRklHStpnqR5q1atqlO5ZmZ9Q0OuKQATIqJV0lbAbZIerXTGiLgUuBRg3LhxUasCzcz6ooYcKUREa/r7DHATMB5YKWkYQPr7TCNqMzPry+oeCpIGStq8bRjYB1gIzAampcmmAb+sd21mZn1dI04fbQ3cJKlt/VdHxG8lPQDMknQ0sAw4pAG1mZn1aXUPhYj4K/BPZdpXA5+sdz1mZvamnvSRVDMzazCHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVnOoWBmZjmHgpmZ5RwKZmaWcyiYmVmux4WCpMmSlkh6XNKpja7HzKwv6VGhIGkj4P8BnwJ2BD4jacfGVmVm1nf0qFAAxgOPR8RfI+IfwLXAlAbXZGbWZygiGl1DTtLBwOSIOCbd/zywW0ScWJjmWODYdHcHYEkdShsC/K0O66k396t3cb96l57cr/dGxNByI/rVu5JOqEzbW1IrIi4FLq1PORlJ8yJiXD3XWQ/uV+/ifvUuvbVfPe30UQvQVLg/EmhtUC1mZn1OTwuFB4DRkraVtAkwFZjd4JrMzPqMHnX6KCLWSToR+B2wETAzIhY1uCyo8+mqOnK/ehf3q3fplf3qUReazcyssXra6SMzM2sgh4KZmeUcCoCkLSTdJumx9HdwO9PNlPSMpIVlxn0pfT3HIkln177qznW3X5KmS3pK0vx027c+lXesGo9XGv81SSFpSG0rrlwVHrPvS1qQHq9bJQ2vT+Udq0K/fijp0dS3myS9uy6Fd6IK/TokvWa8IalHfHzVoZA5FZgTEaOBOel+OZcDk0sbJX2C7D+vmyNiJ+CcGtXZVd3qV3J+RIxNt1tqUOP66Ha/JDUB/wwsq0WB3dDdvv0wIpojYizwa+DbtShyPXS3X7cBO0dEM/D/gdNqUeR66G6/FgIHAXfXpLr1ERF9/kb2X9HD0vAwYEkH044CFpa0zQL2bnQ/atCv6cDXGt2Pavcrtf8C+CdgKTCk0X2qZt8K408DLmp0n2rQrwOBqxrdp2r2C7gTGNfo/kSEjxSSrSNiBUD6u1UX598e2EPSfZLukrRr1StcP93tF8CJ6ZB9ZnuHxg3QrX5J2h94KiL+XIviuqnbj5mkMyUtBz5HzzlSqMa+2OYo4DdVqar7qtmvHqFH/Z9CLUm6HXhPmVHfqsLi+wGDgY8AuwKzJL0v0luAWqpxvy4Cvk/2VSPfB84le0LWXK36JWlAWsY+3VlON2uo5WNGRHwL+Jak04ATge9UY7mdqXW/0jq+BawDrqrWMitYZ8371ZP0mVCIiL3bGydppaRhEbFC0jDgmS4uvgW4MYXA/ZLeIPsyrFXrX3FlatmviFhZWNZPyM5R10UN+/V+YFvgz5Ig+yqVhySNj4inu1V0hWq8LxZdDfwXdQqFWvdL0jRgP+CT9XjD1aaOj1eP4NNHmdnAtDQ8DfhlF+e/GdgLQNL2wCb0jG9H7Fa/0k7e5kCyi2I9wXr3KyIejoitImJURIwiC/Rd6hUIFejuYza6cHd/4NEq1dVd3e3XZOCbwP4R8VKVa+uO7r529DyNvqjRE27AlmSfHHgs/d0itQ8HbilMdw2wAniN7MXk6NS+CfBzshfNh4C9Gt2nKvXrZ8DDwAKynX9Yo/tUjX6VLGspPetCc3cfsxvSfrgA+BUwotF9qlK/HgeWA/PT7eJG96lK/Tow3X8VWAn8rtF98tdcmJlZzqePzMws51AwM7OcQ8HMzHIOBTMzyzkUzMws51AwM7OcQ8HMzHL/A10mIwOcBoTWAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sv_subregions_best_estimator = sv_subregions_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "sv_cv_scores_subregions = model_selection.cross_val_score(\n",
    "                    sv_subregions_best_estimator,\n",
    "                    X_train_subregions,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "train_pred_subregions_sv = sv_subregions_best_estimator.predict(X_train_subregions)\n",
    "sns.histplot(train_pred_subregions_sv).set(title='Predicted vs observed - training data - subregions')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1f4d497e",
   "metadata": {},
   "source": [
    "#### SV: with features - metrics - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "id": "f6fd038f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.070\n",
      "Train MAE:  0.057\n",
      "Test MAE:  0.058\n",
      "Train R-squared:  0.034\n",
      "Test R-squared:  0.039\n",
      "Train RMSE:  0.100\n",
      "Test RMSE:  0.081\n"
     ]
    }
   ],
   "source": [
    "test_pred_sv_subregions = sv_subregions_best_estimator.predict(X_test_subregions)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(sv_cv_scores_subregions): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=train_pred_subregions_sv): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=test_pred_sv_subregions): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=train_pred_subregions_sv): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=test_pred_sv_subregions): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=train_pred_subregions_sv, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=test_pred_sv_subregions, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f3d15305",
   "metadata": {},
   "source": [
    "#### SV: with features - plotting observed vs predicted - predictor: subregions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "id": "aa5973c6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))\n",
    "\n",
    "ax1.scatter(y_test, test_pred_sv_subregions, c=\"orange\")\n",
    "\n",
    "p1 = max(max(test_pred_sv_subregions), max(y_test))\n",
    "p2 = min(min(test_pred_sv_subregions), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Support Vector regression on subregions - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(test_pred_sv_subregions), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title=\"Support Vector regression on subregions - natural scale\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "28d26305",
   "metadata": {},
   "source": [
    "#### SV: with features - performance - predictor: subregions\n",
    "\n",
    "Similar to the support vector regression model with region data, subregion data only performs well on the R-squared metric, but performs the same as the naive model for MAE and RMSE."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a1e0b897",
   "metadata": {},
   "source": [
    "#### SV: with features - fit model - predictor: all pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "id": "2f5af9dc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n",
      "Elapsed time: 13.18 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "sv_all_pests_mod = sv_baseline_grid.fit(X_train_all_pests, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6661f46c",
   "metadata": {},
   "source": [
    "#### SV: with features - CV score distribution - predictor: all pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "id": "47d60bcf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.10 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f9459bae7f0>"
      ]
     },
     "execution_count": 178,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sv_all_pests_best_estimator = sv_all_pests_mod.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "sv_cv_scores_all_pests = model_selection.cross_val_score(\n",
    "                    sv_all_pests_best_estimator,\n",
    "                    X_train_all_pests,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "train_pred_all_pests_sv = sv_all_pests_best_estimator.predict(X_train_all_pests)\n",
    "sns.histplot(train_pred_all_pests_sv).set(title='Predicted vs observed - training data - all pests')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f0d0e4c4",
   "metadata": {},
   "source": [
    "#### SV: with features - metrics - predictor: all pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "id": "8259e14b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.068\n",
      "Train MAE:  0.059\n",
      "Test MAE:  0.058\n",
      "Train R-squared:  0.087\n",
      "Test R-squared:  0.056\n",
      "Train RMSE:  0.097\n",
      "Test RMSE:  0.081\n"
     ]
    }
   ],
   "source": [
    "test_pred_sv_all_pests = sv_all_pests_best_estimator.predict(X_test_all_pests)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(sv_cv_scores_all_pests): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=train_pred_all_pests_sv): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=test_pred_sv_all_pests): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=train_pred_all_pests_sv): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=test_pred_sv_all_pests): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=train_pred_all_pests_sv, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=test_pred_sv_all_pests, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2fdaa677",
   "metadata": {},
   "source": [
    "#### SV: with features - plotting observed vs predicted - predictor: all pests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "id": "cf9a2d47",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))\n",
    "\n",
    "ax1.scatter(y_test, test_pred_sv_all_pests, c=\"orange\")\n",
    "\n",
    "p1 = max(max(test_pred_sv_all_pests), max(y_test))\n",
    "p2 = min(min(test_pred_sv_all_pests), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Support Vector regression on pests - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(test_pred_sv_all_pests), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title=\"Support Vector regression on pests - natural scale\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "73a31c77",
   "metadata": {},
   "source": [
    "#### SV: with features - performance - predictor: pests\n",
    "\n",
    "The support vector model performs worse than the baseline on pests for MAE, better for the R-squared metric, and slightly better for RMSE. The pest model had some of the best performance in the other two regression models, so I would have expected better results here as well. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec5194e8",
   "metadata": {},
   "source": [
    "#### SV: with features - fit model - predictor: mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "id": "e111a560",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n",
      "Elapsed time: 5.86 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "sv_mites_regions = sv_baseline_grid.fit(X_train_mites_regions, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7ad179f0",
   "metadata": {},
   "source": [
    "#### SV: with features - CV score distribution - predictor: mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "id": "d93044a6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.06 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f94826b7430>"
      ]
     },
     "execution_count": 182,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sv_mites_regions_best_estimator = sv_mites_regions.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "sv_cv_scores_mites_regions = model_selection.cross_val_score(\n",
    "                    sv_mites_regions_best_estimator,\n",
    "                    X_train_mites_regions,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "train_pred_sv_mites_regions = sv_mites_regions_best_estimator.predict(X_train_mites_regions)\n",
    "sns.histplot(train_pred_sv_mites_regions).set(title='Predicted vs observed - training data - mites and regions')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c336bfa5",
   "metadata": {},
   "source": [
    "#### SV: with features - metrics - predictor: mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "id": "335cc852",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.139\n",
      "Train MAE:  0.056\n",
      "Test MAE:  0.059\n",
      "Train R-squared:  0.033\n",
      "Test R-squared:  0.008\n",
      "Train RMSE:  0.100\n",
      "Test RMSE:  0.083\n"
     ]
    }
   ],
   "source": [
    "test_pred_sv_mites_regions = sv_mites_regions_best_estimator.predict(X_test_mites_regions)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(test_pred_sv_mites_regions): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=train_pred_sv_mites_regions): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=test_pred_sv_mites_regions): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=train_pred_sv_mites_regions): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=test_pred_sv_mites_regions): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=train_pred_sv_mites_regions, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=test_pred_sv_mites_regions, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2bffb6b4",
   "metadata": {},
   "source": [
    "#### SV: with features - plotting observed vs predicted - predictor: mites and regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "id": "5a311b37",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 4))\n",
    "\n",
    "ax1.scatter(y_test, test_pred_sv_mites_regions, c=\"orange\")\n",
    "\n",
    "p1 = max(max(test_pred_sv_mites_regions), max(y_test))\n",
    "p2 = min(min(test_pred_sv_mites_regions), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Support Vector on mites and regions - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(test_pred_sv_mites_regions), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title=\"Support Vector on mites and regions - natural scale\")\n",
    "ax2.axis('equal')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "932b98aa",
   "metadata": {},
   "source": [
    "#### SV: with features - performance - predictor: mites and regions\n",
    "\n",
    "Like the region and subregion models, adding varroa mite data did not make a significant improvement to the quality of results. In fact, R-squared was slightly worse. Further, the difference in data between train and test sets show that this isn't a well-aligned model."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3c98c24a",
   "metadata": {},
   "source": [
    "#### SV: with features - fit model - predictor: Patsy matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "id": "96edb737",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n",
      "Elapsed time: 17.49 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "sv_matrix_train = sv_baseline_grid.fit(train_model_matrix, y_train)\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e186fc64",
   "metadata": {},
   "source": [
    "#### SV: with features - CV score distribution - predictor: Patsy matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "id": "51216290",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 0.04 seconds\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f943965ca60>"
      ]
     },
     "execution_count": 186,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sv_matrix_best_estimator = sv_matrix_train.best_estimator_\n",
    "\n",
    "# Distribution of cross-val score\n",
    "start = time.time()\n",
    "sv_cv_scores_matrix = model_selection.cross_val_score(\n",
    "                    sv_matrix_best_estimator,\n",
    "                    train_model_matrix,\n",
    "                    y_train,\n",
    "                    scoring='neg_mean_absolute_error',\n",
    "                    cv=cross_validator,\n",
    "                )\n",
    "end = time.time()\n",
    "print(f\"Elapsed time: {end - start:.2f} seconds\")\n",
    "\n",
    "sv_train_pred_matrix = sv_matrix_best_estimator.predict(train_model_matrix)\n",
    "sns.histplot(sv_train_pred_matrix).set(title='Predicted vs observed - training data - Patsy matrix')\n",
    "plt.axvline(x=np.mean(y_train), ymin=0, ymax=12, label='observed', color='green')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c296945",
   "metadata": {},
   "source": [
    "#### SV: with features - metrics - predictor: Patsy matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "id": "b2eacda7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean train CV score: -0.070\n",
      "Train MAE:  0.058\n",
      "Test MAE:  0.057\n",
      "Train R-squared: -0.003\n",
      "Test R-squared: -0.007\n",
      "Train RMSE:  0.102\n",
      "Test RMSE:  0.083\n"
     ]
    }
   ],
   "source": [
    "sv_test_pred_matrix = sv_matrix_best_estimator.predict(test_model_matrix)\n",
    "\n",
    "print(f\"Mean train CV score: {np.mean(sv_cv_scores_matrix): .3f}\")\n",
    "\n",
    "print(f\"Train MAE: {metrics.median_absolute_error(y_true=y_train, y_pred=sv_train_pred_matrix): .3f}\")\n",
    "print(f\"Test MAE: {metrics.median_absolute_error(y_true=y_test, y_pred=sv_test_pred_matrix): .3f}\")\n",
    "print(f\"Train R-squared: {metrics.r2_score(y_true=y_train, y_pred=sv_train_pred_matrix): .3f}\")\n",
    "print(f\"Test R-squared: {metrics.r2_score(y_true=y_test, y_pred=sv_test_pred_matrix): .3f}\")\n",
    "print(f\"Train RMSE: {metrics.mean_squared_error(y_true=y_train, y_pred=sv_train_pred_matrix, squared=False): .3f}\")\n",
    "print(f\"Test RMSE: {metrics.mean_squared_error(y_true=y_test, y_pred=sv_test_pred_matrix, squared=False): .3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ab49cc71",
   "metadata": {},
   "source": [
    "#### SV: with features - plotting observed vs predicted - predictor: Patsy matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "id": "d744f00c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 4))\n",
    "\n",
    "ax1.scatter(y_test, sv_test_pred_matrix, c=\"orange\")\n",
    "\n",
    "p1 = max(max(sv_test_pred_matrix), max(y_test))\n",
    "p2 = min(min(sv_test_pred_matrix), min(y_test))\n",
    "ax1.plot([p1, p2], [p1, p2], 'b-')\n",
    "ax1.set(xlabel='True colony loss counts - log', ylabel='Predicted', title=\"Support Vector on Patsy matrix - log scale\")\n",
    "\n",
    "ax2.scatter(np.exp(y_test), np.exp(sv_test_pred_matrix), c=\"orange\")\n",
    "ax2.set(xlabel='True colony loss counts', title=\"Support Vector on Patsy matrix - natural scale\")\n",
    "ax2.axis('equal')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b21878c",
   "metadata": {},
   "source": [
    "#### SV: with features - performance - predictor: Patsy matrix\n",
    "\n",
    "The metrics for a Patsy matrix of varroa mites, pesticides, and region data was almost identical to the metrics yielded by the baseline model. In fact, they are so similar that I doublechecked that I didn't have an entry error."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ebbe8445",
   "metadata": {},
   "source": [
    "### Support Vector regression - conclusions\n",
    "\n",
    "Overall, the support vector regression model performed poorly. I don't know if poor performance is why support vector is seldom in use anymore, but as a novice to data science, I found it surprising how clear it is that this regression model is lower quality as compared to the other two."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}