Filon.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import keras as k\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import preprocessing\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dropout,Input, Dense,merge, Add\n",
    "from keras import optimizers\n",
    "from keras.models import Model\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "random_ = np.loadtxt(\"./Filon/random.txt\")\n",
    "green_ = np.loadtxt(\"./Filon/green.txt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pd.DataFrame(columns=[\"amu\",\"beta\"])\n",
    "X[\"amu\"] = random_[:,0]\n",
    "X[\"beta\"] = random_[:,1]\n",
    "Y = pd.DataFrame(columns=[\"G\",\"damu\",\"dbeta\"])\n",
    "Y[\"G\"] = green_[:,0]\n",
    "Y[\"damu\"] = green_[:,1]\n",
    "Y[\"dbeta\"] = green_[:,2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.30, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# scaler = Pre()\n",
    "random_scale = preprocessing.MinMaxScaler([0,5])\n",
    "green_scale = preprocessing.MinMaxScaler([0,5])\n",
    "\n",
    "random_scale.fit(X_train)\n",
    "green_scale.fit(y_train)\n",
    "X_train = random_scale.transform(X_train)\n",
    "X_test = random_scale.transform(X_test)\n",
    "\n",
    "y_train = green_scale.transform(y_train)\n",
    "y_test = green_scale.transform(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Y1,Y2,Y3 =y_train[:,0],y_train[:,1],y_train[:,2]\n",
    "# Y1_,Y2_,Y3_ =y_test[:,0],y_test[:,1],y_test[:,2]\n",
    "Y1,Y2,Y3 =y_train[\"G\"],y_train[\"damu\"],y_train[\"dbeta\"]\n",
    "Y1_,Y2_,Y3_ =y_test[\"G\"],y_test[\"damu\"],y_test[\"dbeta\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y1,Y2,Y3 = Y1.reshape(Y1.size,1),Y2.reshape(Y2.size,1),Y3.reshape(Y3.size,1)\n",
    "Y1_,Y2_,Y3_ = Y1_.reshape(Y1_.size,1),Y2_.reshape(Y2_.size,1),Y3_.reshape(Y3_.size,1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model = Sequential()\n",
    "# model.add(Dense(100, input_dim=2, activation='relu'))\n",
    "# model.add(Dense(100, activation='relu'))\n",
    "# model.add(Dense(1000, activation='relu'))\n",
    "# model.add(Dense(1000, activation='relu'))\n",
    "# model.add(Dense(1000, activation='relu'))\n",
    "# model.add(Dense(500, activation='relu'))\n",
    "# model.add(Dense(50, activation='relu'))\n",
    "# model.add(Dense(10, activation='relu'))\n",
    "# model.add(Dropout(0.5))\n",
    "# model.add(Dense(3, activation=None))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "inputs = Input(shape=(2,))\n",
    "x = Dense(8, activation='tanh')(inputs)\n",
    "x = Dense(64, activation='tanh')(x)\n",
    "x = Dense(64, activation='tanh')(x)\n",
    "x = Dense(64, activation='tanh')(x)\n",
    "x_1 = Dense(64, activation='tanh')(x)\n",
    "x = Dense(64, activation='tanh')(x_1)\n",
    "# x = Dense(64, activation='relu')(x)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "x1_ = Dense(64, activation='tanh')(x)\n",
    "x1 = Dense(64, activation='tanh')(x1_)\n",
    "x1 = Dense(64, activation='tanh')(x1)\n",
    "x1 = Dense(64, activation='tanh')(x1)\n",
    "x1 = Dense(64, activation='tanh')(x1)\n",
    "# x = Dense(64, activation='relu')(x)\n",
    "x1 = Add()([x1, x1_])\n",
    "x1 = Dense(64, activation='tanh')(x)\n",
    "y1 = Dense(1, activation=None)(x1)\n",
    "\n",
    "x2_ = Dense(64, activation='tanh')(x)\n",
    "x2 = Dense(64, activation='tanh')(x2_)\n",
    "x2 = Dense(64, activation='tanh')(x2)\n",
    "x2 = Dense(64, activation='tanh')(x2)\n",
    "x2 = Dense(64, activation='tanh')(x2)\n",
    "x2 = Dense(64, activation='tanh')(x2)\n",
    "# x = Dense(64, activation='relu')(x)\n",
    "x2 = Add()([x2, x2_])\n",
    "y2 = Dense(1, activation=None)(x2)\n",
    "\n",
    "x3 = Dense(32, activation='tanh')(x)\n",
    "x3 = Dense(10, activation='tanh')(x3)\n",
    "y3 = Dense(1, activation=None)(x3)\n",
    "\n",
    "model = Model(inputs=inputs, outputs=[y1,y2,y3])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "sgd = optimizers.adam(lr=0.0001)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss=['mean_squared_error','mean_squared_error','mean_squared_error'], optimizer=sgd, metrics=['mae'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 28440 samples, validate on 12189 samples\n",
      "Epoch 1/20\n",
      "28440/28440 [==============================] - 6s 220us/step - loss: 2.5548 - dense_36_loss: 1.0387 - dense_43_loss: 0.5767 - dense_46_loss: 0.9394 - dense_36_mean_absolute_error: 0.6415 - dense_43_mean_absolute_error: 0.4536 - dense_46_mean_absolute_error: 0.5845 - val_loss: 1.4406 - val_dense_36_loss: 0.5951 - val_dense_43_loss: 0.2622 - val_dense_46_loss: 0.5832 - val_dense_36_mean_absolute_error: 0.4911 - val_dense_43_mean_absolute_error: 0.3369 - val_dense_46_mean_absolute_error: 0.4208\n",
      "Epoch 2/20\n",
      "28440/28440 [==============================] - 5s 191us/step - loss: 1.2019 - dense_36_loss: 0.5220 - dense_43_loss: 0.1107 - dense_46_loss: 0.5692 - dense_36_mean_absolute_error: 0.4632 - dense_43_mean_absolute_error: 0.2223 - dense_46_mean_absolute_error: 0.4157 - val_loss: 1.0362 - val_dense_36_loss: 0.4250 - val_dense_43_loss: 0.0559 - val_dense_46_loss: 0.5553 - val_dense_36_mean_absolute_error: 0.4125 - val_dense_43_mean_absolute_error: 0.1467 - val_dense_46_mean_absolute_error: 0.4115\n",
      "Epoch 3/20\n",
      "28440/28440 [==============================] - 5s 191us/step - loss: 0.6663 - dense_36_loss: 0.2086 - dense_43_loss: 0.0383 - dense_46_loss: 0.4194 - dense_36_mean_absolute_error: 0.2994 - dense_43_mean_absolute_error: 0.1270 - dense_46_mean_absolute_error: 0.3729 - val_loss: 0.3295 - val_dense_36_loss: 0.0817 - val_dense_43_loss: 0.0247 - val_dense_46_loss: 0.2231 - val_dense_36_mean_absolute_error: 0.1877 - val_dense_43_mean_absolute_error: 0.1074 - val_dense_46_mean_absolute_error: 0.2907\n",
      "Epoch 4/20\n",
      "28440/28440 [==============================] - 5s 193us/step - loss: 0.1857 - dense_36_loss: 0.0496 - dense_43_loss: 0.0172 - dense_46_loss: 0.1189 - dense_36_mean_absolute_error: 0.1382 - dense_43_mean_absolute_error: 0.0810 - dense_46_mean_absolute_error: 0.2169 - val_loss: 0.1102 - val_dense_36_loss: 0.0267 - val_dense_43_loss: 0.0170 - val_dense_46_loss: 0.0665 - val_dense_36_mean_absolute_error: 0.0922 - val_dense_43_mean_absolute_error: 0.0757 - val_dense_46_mean_absolute_error: 0.1726\n",
      "Epoch 5/20\n",
      "28440/28440 [==============================] - 5s 193us/step - loss: 0.0799 - dense_36_loss: 0.0172 - dense_43_loss: 0.0080 - dense_46_loss: 0.0547 - dense_36_mean_absolute_error: 0.0797 - dense_43_mean_absolute_error: 0.0535 - dense_46_mean_absolute_error: 0.1546 - val_loss: 0.0580 - val_dense_36_loss: 0.0137 - val_dense_43_loss: 0.0047 - val_dense_46_loss: 0.0395 - val_dense_36_mean_absolute_error: 0.0760 - val_dense_43_mean_absolute_error: 0.0401 - val_dense_46_mean_absolute_error: 0.1334\n",
      "Epoch 6/20\n",
      "28440/28440 [==============================] - 6s 194us/step - loss: 0.0450 - dense_36_loss: 0.0094 - dense_43_loss: 0.0035 - dense_46_loss: 0.0321 - dense_36_mean_absolute_error: 0.0617 - dense_43_mean_absolute_error: 0.0360 - dense_46_mean_absolute_error: 0.1176 - val_loss: 0.0325 - val_dense_36_loss: 0.0076 - val_dense_43_loss: 0.0028 - val_dense_46_loss: 0.0222 - val_dense_36_mean_absolute_error: 0.0580 - val_dense_43_mean_absolute_error: 0.0346 - val_dense_46_mean_absolute_error: 0.0935\n",
      "Epoch 7/20\n",
      "28440/28440 [==============================] - 6s 196us/step - loss: 0.0248 - dense_36_loss: 0.0059 - dense_43_loss: 0.0020 - dense_46_loss: 0.0170 - dense_36_mean_absolute_error: 0.0485 - dense_43_mean_absolute_error: 0.0268 - dense_46_mean_absolute_error: 0.0810 - val_loss: 0.0191 - val_dense_36_loss: 0.0055 - val_dense_43_loss: 0.0017 - val_dense_46_loss: 0.0119 - val_dense_36_mean_absolute_error: 0.0443 - val_dense_43_mean_absolute_error: 0.0240 - val_dense_46_mean_absolute_error: 0.0641\n",
      "Epoch 8/20\n",
      "28440/28440 [==============================] - 6s 201us/step - loss: 0.0130 - dense_36_loss: 0.0035 - dense_43_loss: 0.0012 - dense_46_loss: 0.0082 - dense_36_mean_absolute_error: 0.0350 - dense_43_mean_absolute_error: 0.0222 - dense_46_mean_absolute_error: 0.0523 - val_loss: 0.0085 - val_dense_36_loss: 0.0025 - val_dense_43_loss: 7.4510e-04 - val_dense_46_loss: 0.0052 - val_dense_36_mean_absolute_error: 0.0289 - val_dense_43_mean_absolute_error: 0.0205 - val_dense_46_mean_absolute_error: 0.0416\n",
      "Epoch 9/20\n",
      "28440/28440 [==============================] - 6s 202us/step - loss: 0.0076 - dense_36_loss: 0.0022 - dense_43_loss: 7.9534e-04 - dense_46_loss: 0.0046 - dense_36_mean_absolute_error: 0.0270 - dense_43_mean_absolute_error: 0.0185 - dense_46_mean_absolute_error: 0.0408 - val_loss: 0.0056 - val_dense_36_loss: 0.0017 - val_dense_43_loss: 6.7552e-04 - val_dense_46_loss: 0.0033 - val_dense_36_mean_absolute_error: 0.0223 - val_dense_43_mean_absolute_error: 0.0159 - val_dense_46_mean_absolute_error: 0.0361\n",
      "Epoch 10/20\n",
      "28440/28440 [==============================] - 6s 205us/step - loss: 0.0051 - dense_36_loss: 0.0014 - dense_43_loss: 6.5121e-04 - dense_46_loss: 0.0030 - dense_36_mean_absolute_error: 0.0224 - dense_43_mean_absolute_error: 0.0173 - dense_46_mean_absolute_error: 0.0345 - val_loss: 0.0041 - val_dense_36_loss: 0.0012 - val_dense_43_loss: 5.6168e-04 - val_dense_46_loss: 0.0023 - val_dense_36_mean_absolute_error: 0.0224 - val_dense_43_mean_absolute_error: 0.0202 - val_dense_46_mean_absolute_error: 0.0307\n",
      "Epoch 11/20\n",
      "28440/28440 [==============================] - 6s 209us/step - loss: 0.0037 - dense_36_loss: 0.0010 - dense_43_loss: 4.6066e-04 - dense_46_loss: 0.0022 - dense_36_mean_absolute_error: 0.0195 - dense_43_mean_absolute_error: 0.0151 - dense_46_mean_absolute_error: 0.0294 - val_loss: 0.0026 - val_dense_36_loss: 8.7928e-04 - val_dense_43_loss: 1.8606e-04 - val_dense_46_loss: 0.0015 - val_dense_36_mean_absolute_error: 0.0195 - val_dense_43_mean_absolute_error: 0.0086 - val_dense_46_mean_absolute_error: 0.0239\n",
      "Epoch 12/20\n",
      "28440/28440 [==============================] - 6s 210us/step - loss: 0.0029 - dense_36_loss: 7.8837e-04 - dense_43_loss: 5.1523e-04 - dense_46_loss: 0.0015 - dense_36_mean_absolute_error: 0.0175 - dense_43_mean_absolute_error: 0.0154 - dense_46_mean_absolute_error: 0.0238 - val_loss: 0.0021 - val_dense_36_loss: 7.6639e-04 - val_dense_43_loss: 3.0691e-04 - val_dense_46_loss: 9.7837e-04 - val_dense_36_mean_absolute_error: 0.0214 - val_dense_43_mean_absolute_error: 0.0152 - val_dense_46_mean_absolute_error: 0.0177\n",
      "Epoch 13/20\n",
      "28440/28440 [==============================] - 6s 215us/step - loss: 0.0020 - dense_36_loss: 6.1039e-04 - dense_43_loss: 3.3456e-04 - dense_46_loss: 0.0011 - dense_36_mean_absolute_error: 0.0160 - dense_43_mean_absolute_error: 0.0123 - dense_46_mean_absolute_error: 0.0196 - val_loss: 0.0019 - val_dense_36_loss: 5.8054e-04 - val_dense_43_loss: 3.8374e-04 - val_dense_46_loss: 9.0415e-04 - val_dense_36_mean_absolute_error: 0.0150 - val_dense_43_mean_absolute_error: 0.0137 - val_dense_46_mean_absolute_error: 0.0188\n",
      "Epoch 14/20\n",
      "28440/28440 [==============================] - 6s 215us/step - loss: 0.0017 - dense_36_loss: 5.0381e-04 - dense_43_loss: 3.5132e-04 - dense_46_loss: 8.1828e-04 - dense_36_mean_absolute_error: 0.0144 - dense_43_mean_absolute_error: 0.0128 - dense_46_mean_absolute_error: 0.0180 - val_loss: 0.0017 - val_dense_36_loss: 4.7145e-04 - val_dense_43_loss: 4.4575e-04 - val_dense_46_loss: 8.2362e-04 - val_dense_36_mean_absolute_error: 0.0142 - val_dense_43_mean_absolute_error: 0.0145 - val_dense_46_mean_absolute_error: 0.0203\n",
      "Epoch 15/20\n",
      "28440/28440 [==============================] - 7s 232us/step - loss: 0.0014 - dense_36_loss: 4.0685e-04 - dense_43_loss: 2.8815e-04 - dense_46_loss: 6.5661e-04 - dense_36_mean_absolute_error: 0.0129 - dense_43_mean_absolute_error: 0.0116 - dense_46_mean_absolute_error: 0.0165 - val_loss: 8.6780e-04 - val_dense_36_loss: 3.0290e-04 - val_dense_43_loss: 1.2573e-04 - val_dense_46_loss: 4.3917e-04 - val_dense_36_mean_absolute_error: 0.0110 - val_dense_43_mean_absolute_error: 0.0087 - val_dense_46_mean_absolute_error: 0.0137\n",
      "Epoch 16/20\n",
      "28440/28440 [==============================] - 6s 226us/step - loss: 0.0011 - dense_36_loss: 3.4110e-04 - dense_43_loss: 2.7245e-04 - dense_46_loss: 5.2783e-04 - dense_36_mean_absolute_error: 0.0117 - dense_43_mean_absolute_error: 0.0110 - dense_46_mean_absolute_error: 0.0150 - val_loss: 7.0100e-04 - val_dense_36_loss: 2.1904e-04 - val_dense_43_loss: 1.2815e-04 - val_dense_46_loss: 3.5381e-04 - val_dense_36_mean_absolute_error: 0.0093 - val_dense_43_mean_absolute_error: 0.0094 - val_dense_46_mean_absolute_error: 0.0126\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 17/20\n",
      "28440/28440 [==============================] - 6s 217us/step - loss: 0.0010 - dense_36_loss: 3.0698e-04 - dense_43_loss: 2.5280e-04 - dense_46_loss: 4.6222e-04 - dense_36_mean_absolute_error: 0.0111 - dense_43_mean_absolute_error: 0.0108 - dense_46_mean_absolute_error: 0.0141 - val_loss: 6.6725e-04 - val_dense_36_loss: 1.6881e-04 - val_dense_43_loss: 2.1043e-04 - val_dense_46_loss: 2.8801e-04 - val_dense_36_mean_absolute_error: 0.0083 - val_dense_43_mean_absolute_error: 0.0114 - val_dense_46_mean_absolute_error: 0.0111\n",
      "Epoch 18/20\n",
      "28440/28440 [==============================] - 6s 225us/step - loss: 8.4412e-04 - dense_36_loss: 2.6269e-04 - dense_43_loss: 2.0327e-04 - dense_46_loss: 3.7816e-04 - dense_36_mean_absolute_error: 0.0103 - dense_43_mean_absolute_error: 0.0097 - dense_46_mean_absolute_error: 0.0128 - val_loss: 7.9970e-04 - val_dense_36_loss: 2.9333e-04 - val_dense_43_loss: 1.2264e-04 - val_dense_46_loss: 3.8373e-04 - val_dense_36_mean_absolute_error: 0.0099 - val_dense_43_mean_absolute_error: 0.0073 - val_dense_46_mean_absolute_error: 0.0141\n",
      "Epoch 19/20\n",
      "28440/28440 [==============================] - 6s 220us/step - loss: 8.0599e-04 - dense_36_loss: 2.4247e-04 - dense_43_loss: 2.1619e-04 - dense_46_loss: 3.4732e-04 - dense_36_mean_absolute_error: 0.0098 - dense_43_mean_absolute_error: 0.0101 - dense_46_mean_absolute_error: 0.0122 - val_loss: 3.8449e-04 - val_dense_36_loss: 1.2902e-04 - val_dense_43_loss: 5.7433e-05 - val_dense_46_loss: 1.9804e-04 - val_dense_36_mean_absolute_error: 0.0072 - val_dense_43_mean_absolute_error: 0.0049 - val_dense_46_mean_absolute_error: 0.0096\n",
      "Epoch 20/20\n",
      "28440/28440 [==============================] - 6s 221us/step - loss: 6.7030e-04 - dense_36_loss: 2.0543e-04 - dense_43_loss: 1.7737e-04 - dense_46_loss: 2.8751e-04 - dense_36_mean_absolute_error: 0.0089 - dense_43_mean_absolute_error: 0.0089 - dense_46_mean_absolute_error: 0.0112 - val_loss: 7.1784e-04 - val_dense_36_loss: 2.4537e-04 - val_dense_43_loss: 1.5605e-04 - val_dense_46_loss: 3.1642e-04 - val_dense_36_mean_absolute_error: 0.0128 - val_dense_43_mean_absolute_error: 0.0097 - val_dense_46_mean_absolute_error: 0.0114\n"
     ]
    }
   ],
   "source": [
    "history = model.fit([X_train], [Y1,Y2,Y3], epochs=20, batch_size=32,validation_data=([X_test],[Y1_,Y2_,Y3_]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc='upper left')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "input_ = random_scale.transform(X)\n",
    "output_ = model.predict(input_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "output_ = np.transpose(np.array(output_))\n",
    "output_ = green_scale.inverse_transform(output_.reshape(output_.shape[1],3))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f1c75b8e6a0>]"
      ]
     },
     "execution_count": 63,
     "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": [
    "plt.plot(output_[:,2],green_[:,2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.datasets import make_regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=2,\n",
       "                      max_features='auto', max_leaf_nodes=None,\n",
       "                      min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                      min_samples_leaf=1, min_samples_split=2,\n",
       "                      min_weight_fraction_leaf=0.0, n_estimators=100,\n",
       "                      n_jobs=None, oob_score=False, random_state=0, verbose=0,\n",
       "                      warm_start=False)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr = RandomForestRegressor(max_depth=2000, random_state=0,\n",
    "                             n_estimators=1000)\n",
    "regr.fit(X_train, Y1)  \n",
    "RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=2,\n",
    "           max_features='auto', max_leaf_nodes=None,\n",
    "           min_impurity_decrease=0.0, min_impurity_split=None,\n",
    "           min_samples_leaf=1, min_samples_split=2,\n",
    "           min_weight_fraction_leaf=0.0, n_estimators=100, n_jobs=None,\n",
    "           oob_score=False, random_state=0, verbose=0, warm_start=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.047833444514077325"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(np.sum(np.square(regr.predict(X_test)-Y1_)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}