diff --git a/02_activities/assignments/assignment_1.ipynb b/02_activities/assignments/assignment_1.ipynb index 45cfc9cd7..e326588d2 100644 --- a/02_activities/assignments/assignment_1.ipynb +++ b/02_activities/assignments/assignment_1.ipynb @@ -26,11 +26,15 @@ }, { "cell_type": "code", - "execution_count": 2, + "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "# Write your code below.\n", + "\n", + "# Load environment variables from a .env file\n", + "%load_ext dotenv\n", + "%dotenv\n", "\n" ] }, @@ -55,14 +59,34 @@ }, { "cell_type": "code", - "execution_count": 3, + "execution_count": 2, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "PRICE_DATA: ../../05_src/data/prices/\n", + "Number of parquet files found: 3083\n" + ] + } + ], "source": [ "import os\n", "from glob import glob\n", "\n", "# Write your code below.\n", + "import dask.dataframe as dd\n", + "\n", + "# Load the PRICE_DATA environment variable\n", + "PRICE_DATA = os.getenv(\"PRICE_DATA\")\n", + "\n", + "# Use glob to find all parquet files under PRICE_DATA (recursively)\n", + "parquet_files = glob(os.path.join(PRICE_DATA, \"**\", \"*.parquet\"), recursive=True)\n", + "\n", + "# Confirm results\n", + "print(f\"PRICE_DATA: {PRICE_DATA}\")\n", + "print(f\"Number of parquet files found: {len(parquet_files)}\")\n", "\n" ] }, @@ -88,12 +112,103 @@ }, { "cell_type": "code", - "execution_count": 4, + "execution_count": null, "metadata": {}, - "outputs": [], + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
DateOpenHighLowCloseAdj_CloseVolumesourcetickerYearClose_lag_1Adj_Close_lag_1returnshi_lo_range
\n", + "
" + ], + "text/plain": [ + "Empty DataFrame\n", + "Columns: [Date, Open, High, Low, Close, Adj_Close, Volume, source, ticker, Year, Close_lag_1, Adj_Close_lag_1, returns, hi_lo_range]\n", + "Index: []" + ] + }, + "execution_count": 13, + "metadata": {}, + "output_type": "execute_result" + } + ], "source": [ - "# Write your code below.\n", - "\n" + "import numpy as np\n", + "import dask.dataframe as dd\n", + "\n", + "# Assumes `parquet_files` is already defined.\n", + "dd_px = dd.read_parquet(parquet_files)\n", + "\n", + "# Normalize adjusted-close column name if needed\n", + "if \"Adj_Close\" not in dd_px.columns and \"Adj Close\" in dd_px.columns:\n", + " dd_px = dd_px.rename(columns={\"Adj Close\": \"Adj_Close\"})\n", + "\n", + "# Ensure Date is datetime64[ns] (no-op if already correct)\n", + "if dd_px[\"Date\"].dtype.kind not in (\"M\",):\n", + " dd_px = dd_px.assign(Date=dd_px[\"Date\"].astype(\"M8[ns]\"))\n", + "\n", + "# ---- Build explicit meta (schema) for the output with the four new columns ----\n", + "meta = (\n", + " dd_px._meta.assign(\n", + " Close_lag_1=np.float64(),\n", + " Adj_Close_lag_1=np.float64(),\n", + " returns=np.float64(),\n", + " hi_lo_range=np.float64(),\n", + " )\n", + ")\n", + "\n", + "# ---- Per-ticker feature logic: sort by Date, add lags/returns/range ----\n", + "def _per_ticker_features(pdf):\n", + " # NOTE: This function is executed by Dask on each group partition;\n", + " # the returned result is stitched back into a single Dask DataFrame.\n", + " pdf = pdf.sort_values(\"Date\")\n", + " pdf[\"Close_lag_1\"] = pdf[\"Close\"].shift(1)\n", + " pdf[\"Adj_Close_lag_1\"] = pdf[\"Adj_Close\"].shift(1)\n", + " pdf[\"returns\"] = pdf[\"Close\"] / pdf[\"Close_lag_1\"] - 1\n", + " pdf[\"hi_lo_range\"] = pdf[\"High\"] - pdf[\"Low\"]\n", + " return pdf\n", + "\n", + "# Group by ticker and apply the feature function with explicit meta\n", + "dd_feat = dd_px.groupby(\"ticker\", group_keys=False).apply(_per_ticker_features, meta=meta)\n", + "\n", + "# Quick peek \n", + "dd_feat.head()\n" ] }, { @@ -108,11 +223,197 @@ }, { "cell_type": "code", - "execution_count": 25, + "execution_count": null, "metadata": {}, - "outputs": [], + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
indexDateOpenHighLowCloseAdj_CloseVolumesourcetickerYearClose_lag_1Adj_Close_lag_1returnshi_lo_rangereturns_ma10
01373152001-07-1915.1015.2915.0015.1711.40439434994300.0ACN.csvACN2001NaNNaNNaN0.29NaN
11373162001-07-2015.0515.0514.8015.0111.2841089238500.0ACN.csvACN200115.1711.404394-0.0105470.25NaN
21373172001-07-2315.0015.0114.5515.0011.2765877501000.0ACN.csvACN200115.0111.284108-0.0006660.46NaN
31373182001-07-2414.9514.9714.7014.8611.1713413537300.0ACN.csvACN200115.0011.276587-0.0093330.27NaN
41373192001-07-2514.7014.9514.6514.9511.2389994208100.0ACN.csvACN200114.8611.1713410.0060570.30NaN
\n", + "
" + ], + "text/plain": [ + " index Date Open High Low Close Adj_Close Volume \\\n", + "0 137315 2001-07-19 15.10 15.29 15.00 15.17 11.404394 34994300.0 \n", + "1 137316 2001-07-20 15.05 15.05 14.80 15.01 11.284108 9238500.0 \n", + "2 137317 2001-07-23 15.00 15.01 14.55 15.00 11.276587 7501000.0 \n", + "3 137318 2001-07-24 14.95 14.97 14.70 14.86 11.171341 3537300.0 \n", + "4 137319 2001-07-25 14.70 14.95 14.65 14.95 11.238999 4208100.0 \n", + "\n", + " source ticker Year Close_lag_1 Adj_Close_lag_1 returns hi_lo_range \\\n", + "0 ACN.csv ACN 2001 NaN NaN NaN 0.29 \n", + "1 ACN.csv ACN 2001 15.17 11.404394 -0.010547 0.25 \n", + "2 ACN.csv ACN 2001 15.01 11.284108 -0.000666 0.46 \n", + "3 ACN.csv ACN 2001 15.00 11.276587 -0.009333 0.27 \n", + "4 ACN.csv ACN 2001 14.86 11.171341 0.006057 0.30 \n", + "\n", + " returns_ma10 \n", + "0 NaN \n", + "1 NaN \n", + "2 NaN \n", + "3 NaN \n", + "4 NaN " + ] + }, + "execution_count": 14, + "metadata": {}, + "output_type": "execute_result" + } + ], "source": [ "# Write your code below.\n", + "\n", + "\n", + "# Convert the Dask DataFrame to a pandas DataFrame\n", + "# Assumes `dd_feat` exists from the previous step\n", + "pdf_feat = dd_feat.reset_index().compute()\n", + "\n", + "# Ensure chronological order within each ticker\n", + "pdf_feat = pdf_feat.sort_values([\"ticker\", \"Date\"])\n", + "\n", + "# Add a 10-day moving average of returns (per ticker)\n", + "# transform keeps the result aligned to the original rows\n", + "pdf_feat[\"returns_ma10\"] = (\n", + " pdf_feat.groupby(\"ticker\")[\"returns\"]\n", + " .transform(lambda s: s.rolling(window=10, min_periods=10).mean())\n", + ")\n", + "\n", + "# Optional quick check\n", + "pdf_feat.head()\n", + "\n", "\n" ] }, @@ -123,7 +424,24 @@ "Please comment:\n", "\n", "+ Was it necessary to convert to pandas to calculate the moving average return?\n", + "No. Dask can compute a moving average. \n", + "Converting to pandas was a pragmatic choice for simplicity and quick validation, but it isn’t strictly required for this feature.\n", "+ Would it have been better to do it in Dask? Why?\n", + "Yes, in the case where data size is large or you want to stay scalable. \n", + "1. Scalability & memory: \n", + "Dask: Keeps computation lazy and chunked by partitions. \n", + "pandas: Fine for small/medium data, but it breaks scalability and risks memory pressure.\n", + "\n", + "2. Performance & parallelism\n", + "Dask: Rolling operations can be parallelized across partitions. W\n", + "pandas: Runs eagerly on one process; you lose the benefits of the Dask task graph you built up to this point.\n", + "\n", + "3. Operational consistency\n", + "Keeping the whole pipeline in Dask means one execution model (lazy graph → write to Parquet), consistent dtypes/schemas, and easier productionization (no sudden memory spikes from .compute()).\n", + "\n", + "\n", + "\n", + "\n", "\n", "(1 pt)" ] @@ -165,7 +483,7 @@ ], "metadata": { "kernelspec": { - "display_name": "env", + "display_name": "production-env", "language": "python", "name": "python3" }, @@ -179,7 +497,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.11.0" + "version": "3.11.14" } }, "nbformat": 4, diff --git a/02_activities/assignments/assignment_2.ipynb b/02_activities/assignments/assignment_2.ipynb index 29d661c57..bc5971a1a 100644 --- a/02_activities/assignments/assignment_2.ipynb +++ b/02_activities/assignments/assignment_2.ipynb @@ -97,24 +97,122 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 1, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "c:\\Users\\bprabaharan\\Downloads\\Production\\production-env\\Scripts\\python.exe\n" + ] + }, + { + "name": "stderr", + "output_type": "stream", + "text": [ + "c:\\Users\\bprabaharan\\Downloads\\Production\\production-env\\Lib\\site-packages\\tqdm\\auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n", + " from .autonotebook import tqdm as notebook_tqdm\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "SHAP version: 0.50.0\n" + ] + } + ], "source": [ - "# Load the libraries as required." + "import sys\n", + "print(sys.executable)\n", + "import shap\n", + "print(\"SHAP version:\", shap.__version__)\n" ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 3, "metadata": {}, "outputs": [], "source": [ + "# Load the libraries as required.\n", + "\n", + "# Data handling\n", + "import pandas as pd\n", + "import numpy as np\n", + "\n", + "# Preprocessing\n", + "from sklearn.model_selection import train_test_split, GridSearchCV\n", + "from sklearn.compose import ColumnTransformer\n", + "from sklearn.preprocessing import OneHotEncoder, StandardScaler, PolynomialFeatures\n", + "from sklearn.pipeline import Pipeline\n", + "\n", + "# Models\n", + "from sklearn.linear_model import LinearRegression, Ridge, Lasso\n", + "from sklearn.neighbors import KNeighborsRegressor\n", + "from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor\n", + "\n", + "# Metrics\n", + "from sklearn.metrics import mean_squared_error, mean_absolute_error\n", + "\n", + "# Model export\n", + "import pickle\n", + "\n", + "# Explainability\n", + "import shap\n", + "\n", + "# Visualization\n", + "import matplotlib.pyplot as plt\n", + "import seaborn as sns\n" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "RangeIndex: 517 entries, 0 to 516\n", + "Data columns (total 13 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 coord_x 517 non-null int64 \n", + " 1 coord_y 517 non-null int64 \n", + " 2 month 517 non-null object \n", + " 3 day 517 non-null object \n", + " 4 ffmc 517 non-null float64\n", + " 5 dmc 517 non-null float64\n", + " 6 dc 517 non-null float64\n", + " 7 isi 517 non-null float64\n", + " 8 temp 517 non-null float64\n", + " 9 rh 517 non-null int64 \n", + " 10 wind 517 non-null float64\n", + " 11 rain 517 non-null float64\n", + " 12 area 517 non-null float64\n", + "dtypes: float64(8), int64(3), object(2)\n", + "memory usage: 52.6+ KB\n" + ] + } + ], + "source": [ + "\n", "# Load data\n", "columns = [\n", - " 'coord_x', 'coord_y', 'month', 'day', 'ffmc', 'dmc', 'dc', 'isi', 'temp', 'rh', 'wind', 'rain', 'area' \n", + " 'coord_x', 'coord_y', 'month', 'day', 'ffmc', 'dmc', 'dc', 'isi',\n", + " 'temp', 'rh', 'wind', 'rain', 'area'\n", "]\n", - "fires_dt = (pd.read_csv('../../05_src/data/fires/forestfires.csv', header = 0, names = columns))\n", + "\n", + "fires_dt = pd.read_csv(\n", + " '../../05_src/data/forest+fires/forestfires.csv',\n", + " header=0,\n", + " names=columns\n", + ")\n", + "\n", "fires_dt.info()\n" ] }, @@ -129,17 +227,59 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 13, "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "X shape: (517, 12)\n", + "y shape: (517,)\n", + "Categorical dtypes (will be encoded later): {'month': dtype('O'), 'day': dtype('O')}\n" + ] + }, + { + "data": { + "text/plain": [ + "((413, 12), (104, 12), (413,), (104,))" + ] + }, + "execution_count": 13, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "\n", + "# 1) Define feature and target columns\n", + "target_col = \"area\"\n", + "feature_cols = [\n", + " 'coord_x', 'coord_y', 'month', 'day',\n", + " 'ffmc', 'dmc', 'dc', 'isi',\n", + " 'temp', 'rh', 'wind', 'rain'\n", + "]\n", + "\n", + "# 2) Create X (features) and y (target)\n", + "X = fires_dt[feature_cols].copy()\n", + "y = fires_dt[target_col].copy()\n", + "\n", + "# Hygiene checks\n", + "print(\"X shape:\", X.shape)\n", + "print(\"y shape:\", y.shape)\n", + "print(\"Categorical dtypes (will be encoded later):\", X[['month','day']].dtypes.to_dict())\n", + "\n", + "from sklearn.model_selection import train_test_split\n", + "\n", + "X_train, X_test, y_train, y_test = train_test_split(\n", + " X, y,\n", + " test_size=0.20,\n", + " random_state=42,\n", + " shuffle=True\n", + ")\n", + "\n", + "X_train.shape, X_test.shape, y_train.shape, y_test.shape\n" + ] }, { "cell_type": "markdown", @@ -180,10 +320,58 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 14, "metadata": {}, - "outputs": [], - "source": [] + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Numeric features: ['coord_x', 'coord_y', 'ffmc', 'dmc', 'dc', 'isi', 'temp', 'rh', 'wind', 'rain']\n", + "Categorical features: ['month', 'day']\n", + "preproc1 is ready.\n" + ] + } + ], + "source": [ + "\n", + "from sklearn.compose import ColumnTransformer\n", + "from sklearn.preprocessing import OneHotEncoder, StandardScaler\n", + "from sklearn.pipeline import Pipeline\n", + "import numpy as np\n", + "import pandas as pd\n", + "\n", + "# 1) Identify column types from X\n", + "numeric_features = X.select_dtypes(include=[np.number]).columns.tolist()\n", + "categorical_features = X.select_dtypes(include=[\"object\", \"category\"]).columns.tolist()\n", + "\n", + "print(\"Numeric features:\", numeric_features)\n", + "print(\"Categorical features:\", categorical_features)\n", + "\n", + "# 2) Build the numeric and categorical pipelines\n", + "numeric_pipe_1 = Pipeline(steps=[\n", + " (\"scaler\", StandardScaler()) # StandardScaler to standardize numeric features\n", + "])\n", + "\n", + "categorical_pipe_1 = Pipeline(steps=[\n", + " (\"ohe\", OneHotEncoder(handle_unknown=\"ignore\", sparse_output=False))\n", + " # handle_unknown=\"ignore\" so inference doesn't break on unseen categories\n", + "])\n", + "\n", + "# 3) Assemble ColumnTransformer (preproc1)\n", + "preproc1 = ColumnTransformer(\n", + " transformers=[\n", + " (\"num\", numeric_pipe_1, numeric_features),\n", + " (\"cat\", categorical_pipe_1, categorical_features),\n", + " ],\n", + " remainder=\"drop\", # keep only the specified columns\n", + " verbose_feature_names_out=True\n", + ")\n", + "\n", + "# Test: fit_transform a small sample to confirm it works\n", + "_ = preproc1.fit(X) # validate the configuration; pipeline fitting occurs later\n", + "print(\"preproc1 is ready.\")\n" + ] }, { "cell_type": "markdown", @@ -191,7 +379,7 @@ "source": [ "### Preproc 2\n", "\n", - "Create preproc1 below.\n", + "Create preproc2 below.\n", "\n", "+ Numeric: scaled variables, non-linear transformation to one or more variables.\n", "+ Categorical: one-hot encoding." @@ -199,10 +387,54 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 10, "metadata": {}, - "outputs": [], - "source": [] + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "preproc2 is ready.\n" + ] + } + ], + "source": [ + "from sklearn.preprocessing import QuantileTransformer\n", + "\n", + "# Reuse the same detected columns\n", + "numeric_features = X.select_dtypes(include=[np.number]).columns.tolist()\n", + "categorical_features = X.select_dtypes(include=[\"object\", \"category\"]).columns.tolist()\n", + "\n", + "# Numeric pipeline with a non-linear transformation\n", + "numeric_pipe_2 = Pipeline(steps=[\n", + " (\"quantile\", QuantileTransformer(\n", + " n_quantiles=min(1000, len(X)), # cap by sample size to avoid warnings\n", + " output_distribution=\"normal\",\n", + " subsample=int(1e9), # disable internal subsampling for reproducibility on small data\n", + " random_state=42\n", + " )),\n", + " (\"scaler\", StandardScaler())\n", + "])\n", + "\n", + "# Categorical pipeline (same as preproc1)\n", + "categorical_pipe_2 = Pipeline(steps=[\n", + " (\"ohe\", OneHotEncoder(handle_unknown=\"ignore\", sparse_output=False))\n", + "])\n", + "\n", + "# Assemble ColumnTransformer (preproc2)\n", + "preproc2 = ColumnTransformer(\n", + " transformers=[\n", + " (\"num\", numeric_pipe_2, numeric_features),\n", + " (\"cat\", categorical_pipe_2, categorical_features),\n", + " ],\n", + " remainder=\"drop\",\n", + " verbose_feature_names_out=True\n", + ")\n", + "\n", + "# test: quick fit check\n", + "_ = preproc2.fit(X)\n", + "print(\"preproc2 is ready.\")\n" + ] }, { "cell_type": "markdown", @@ -229,38 +461,366 @@ "cell_type": "code", "execution_count": null, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "=== Pipeline A: preproc1 + Ridge (Baseline) ===\n", + "Best CV RMSE: 39.4253\n", + "Test RMSE : 108.3514\n", + "Test MAE : 24.1778\n", + "Best Params : {'regressor__alpha': 100.0, 'regressor__fit_intercept': True}\n" + ] + } + ], "source": [ - "# Pipeline A = preproc1 + baseline\n" + "# Pipeline A = preproc1 + baseline\n", + "from sklearn.pipeline import Pipeline\n", + "from sklearn.linear_model import Ridge\n", + "from sklearn.model_selection import GridSearchCV\n", + "from sklearn.metrics import mean_squared_error, mean_absolute_error\n", + "import numpy as np\n", + "import pandas as pd\n", + "\n", + "RANDOM_STATE = 42\n", + "SCORING = \"neg_root_mean_squared_error\" # use RMSE as the selection metric (negated in sklearn)\n", + "CV = 5\n", + "N_JOBS = -1\n", + "\n", + "# -------------------------------\n", + "# Pipeline A (Baseline)\n", + "# preproc1 + Ridge\n", + "# -------------------------------\n", + "pipeA = Pipeline(steps=[\n", + " (\"preprocessing\", preproc1),\n", + " (\"regressor\", Ridge()) # Ridge is our baseline\n", + "])\n", + "\n", + "# A small, reasonable grid (≥ 4 combos) for Ridge\n", + "param_grid_A = {\n", + " \"regressor__alpha\": [0.01, 0.1, 1.0, 10.0, 100.0],\n", + " \"regressor__fit_intercept\": [True, False]\n", + "}\n", + "\n", + "gridA = GridSearchCV(\n", + " estimator=pipeA,\n", + " param_grid=param_grid_A,\n", + " scoring=SCORING,\n", + " cv=CV,\n", + " n_jobs=N_JOBS,\n", + " refit=True,\n", + " verbose=0,\n", + " return_train_score=False\n", + ")\n", + "\n", + "# Fit on training data\n", + "gridA.fit(X_train, y_train)\n", + "\n", + "# Cross-validated best score (convert negative RMSE -> positive RMSE)\n", + "best_cv_rmse_A = -gridA.best_score_\n", + "\n", + "# Hold-out test evaluation\n", + "y_pred_A = gridA.best_estimator_.predict(X_test)\n", + "test_rmse_A = np.sqrt(mean_squared_error(y_test, y_pred_A))\n", + "test_mae_A = mean_absolute_error(y_test, y_pred_A)\n", + "\n", + "print(\"=== Pipeline A: preproc1 + Ridge (Baseline) ===\")\n", + "print(\"Best CV RMSE: \", round(best_cv_rmse_A, 4))\n", + "print(\"Test RMSE : \", round(test_rmse_A, 4))\n", + "print(\"Test MAE : \", round(test_mae_A, 4))\n", + "print(\"Best Params : \", gridA.best_params_)\n", + "\n", + "# Keep the fitted best pipeline handy for later comparison/pickling/SHAP (if it wins)\n", + "best_pipe_A = gridA.best_estimator_\n" ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 19, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "=== Pipeline B: preproc2 + Lasso (Baseline) ===\n", + "Best CV RMSE: 38.7846\n", + "Test RMSE : 108.6361\n", + "Test MAE : 24.3781\n", + "Best Params: {'regressor__alpha': 1.0, 'regressor__fit_intercept': True}\n" + ] + }, + { + "name": "stderr", + "output_type": "stream", + "text": [ + "c:\\Users\\bprabaharan\\Downloads\\Production\\production-env\\Lib\\site-packages\\sklearn\\preprocessing\\_data.py:2846: UserWarning: n_quantiles (517) is greater than the total number of samples (413). n_quantiles is set to n_samples.\n", + " warnings.warn(\n" + ] + } + ], "source": [ - "# Pipeline B = preproc2 + baseline\n" + "# Pipeline B = preproc2 + baseline\n", + "\n", + "# --- Imports ---\n", + "from sklearn.pipeline import Pipeline\n", + "from sklearn.linear_model import Lasso\n", + "from sklearn.model_selection import GridSearchCV\n", + "from sklearn.metrics import mean_squared_error, mean_absolute_error\n", + "from sklearn.compose import ColumnTransformer\n", + "from sklearn.preprocessing import OneHotEncoder, StandardScaler, QuantileTransformer\n", + "import numpy as np\n", + "import pandas as pd\n", + "\n", + "RANDOM_STATE = 42\n", + "SCORING = \"neg_root_mean_squared_error\" # RMSE as selection metric (negated for sklearn)\n", + "CV = 5\n", + "N_JOBS = -1\n", + "\n", + "# --- Ensure preproc2 exists (rebuild only if missing in this kernel) ---\n", + "try:\n", + " preproc2 # noqa: F821\n", + "except NameError:\n", + " numeric_features = X.select_dtypes(include=[np.number]).columns.tolist()\n", + " categorical_features = X.select_dtypes(include=[\"object\", \"category\"]).columns.tolist()\n", + "\n", + " numeric_pipe_2 = Pipeline(steps=[\n", + " (\"quantile\", QuantileTransformer(\n", + " n_quantiles=min(1000, len(X)),\n", + " output_distribution=\"normal\",\n", + " subsample=int(1e9), # disables internal subsampling on small data\n", + " random_state=RANDOM_STATE\n", + " )),\n", + " (\"scaler\", StandardScaler())\n", + " ])\n", + "\n", + " categorical_pipe_2 = Pipeline(steps=[\n", + " (\"ohe\", OneHotEncoder(handle_unknown=\"ignore\", sparse_output=False))\n", + " ])\n", + "\n", + " preproc2 = ColumnTransformer(\n", + " transformers=[\n", + " (\"num\", numeric_pipe_2, numeric_features),\n", + " (\"cat\", categorical_pipe_2, categorical_features),\n", + " ],\n", + " remainder=\"drop\",\n", + " verbose_feature_names_out=True\n", + " )\n", + " _ = preproc2.fit(X)\n", + " print(\"preproc2 was not found and has been created.\")\n", + "\n", + "# --- Build Pipeline B (preproc2 + Lasso) ---\n", + "pipeB = Pipeline(steps=[\n", + " (\"preprocessing\", preproc2),\n", + " (\"regressor\", Lasso(max_iter=10000, random_state=RANDOM_STATE))\n", + "])\n", + "\n", + "# --- Reasonable, diverse hyperparameter grid (>= 4 combos) ---\n", + "param_grid_B = {\n", + " \"regressor__alpha\": [0.0001, 0.0005, 0.001, 0.01, 0.1, 1.0],\n", + " \"regressor__fit_intercept\": [True, False],\n", + "}\n", + "\n", + "gridB = GridSearchCV(\n", + " estimator=pipeB,\n", + " param_grid=param_grid_B,\n", + " scoring=SCORING,\n", + " cv=CV,\n", + " n_jobs=N_JOBS,\n", + " refit=True,\n", + " verbose=0,\n", + " return_train_score=False\n", + ")\n", + "\n", + "# --- Fit on training data ---\n", + "gridB.fit(X_train, y_train)\n", + "\n", + "# --- Evaluation ---\n", + "best_cv_rmse_B = -gridB.best_score_ # convert from neg RMSE to RMSE\n", + "\n", + "y_pred_B = gridB.best_estimator_.predict(X_test)\n", + "# Use version-agnostic RMSE (some sklearn versions don't support squared=False)\n", + "test_rmse_B = np.sqrt(mean_squared_error(y_test, y_pred_B))\n", + "test_mae_B = mean_absolute_error(y_test, y_pred_B)\n", + "\n", + "print(\"=== Pipeline B: preproc2 + Lasso (Baseline) ===\")\n", + "print(\"Best CV RMSE:\", round(best_cv_rmse_B, 4))\n", + "print(\"Test RMSE :\", round(test_rmse_B, 4))\n", + "print(\"Test MAE :\", round(test_mae_B, 4))\n", + "print(\"Best Params:\", gridB.best_params_)\n", + "\n", + "# Keep the fitted best pipeline for later comparison/pickling/SHAP (if it wins)\n", + "best_pipe_B = gridB.best_estimator_\n" ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 20, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "=== Pipeline C: preproc1 + GradientBoostingRegressor (Advanced) ===\n", + "Best CV RMSE: 45.2116\n", + "Test RMSE : 109.3591\n", + "Test MAE : 25.2061\n", + "Best Params: {'regressor__learning_rate': 0.01, 'regressor__max_depth': 2, 'regressor__n_estimators': 100, 'regressor__subsample': 0.8}\n" + ] + } + ], "source": [ - "# Pipeline C = preproc1 + advanced model\n" + "# Pipeline C = preproc1 + advanced model\n", + "\n", + "# === Pipeline C: preproc1 + GradientBoostingRegressor (Advanced) ===\n", + "\n", + "from sklearn.pipeline import Pipeline\n", + "from sklearn.ensemble import GradientBoostingRegressor\n", + "from sklearn.model_selection import GridSearchCV\n", + "from sklearn.metrics import mean_squared_error, mean_absolute_error\n", + "import numpy as np\n", + "\n", + "RANDOM_STATE = 42\n", + "SCORING = \"neg_root_mean_squared_error\" # RMSE (negated for sklearn)\n", + "CV = 5\n", + "N_JOBS = -1\n", + "\n", + "# Build Pipeline C\n", + "pipeC = Pipeline(steps=[\n", + " (\"preprocessing\", preproc1), # scale nums + OHE cats (no non-linear transform)\n", + " (\"regressor\", GradientBoostingRegressor(random_state=RANDOM_STATE))\n", + "])\n", + "\n", + "# Diverse, reasonable grid (>= 4 combos)\n", + "param_grid_C = {\n", + " \"regressor__n_estimators\": [100, 300, 600],\n", + " \"regressor__learning_rate\": [0.01, 0.05, 0.1],\n", + " \"regressor__max_depth\": [2, 3, 4],\n", + " \"regressor__subsample\": [0.8, 1.0],\n", + "}\n", + "\n", + "gridC = GridSearchCV(\n", + " estimator=pipeC,\n", + " param_grid=param_grid_C,\n", + " scoring=SCORING,\n", + " cv=CV,\n", + " n_jobs=N_JOBS,\n", + " refit=True,\n", + " verbose=0,\n", + " return_train_score=False,\n", + ")\n", + "\n", + "# Fit on training data\n", + "gridC.fit(X_train, y_train)\n", + "\n", + "# Cross-validated best score (convert from neg RMSE to RMSE)\n", + "best_cv_rmse_C = -gridC.best_score_\n", + "\n", + "# Hold-out test evaluation\n", + "y_pred_C = gridC.best_estimator_.predict(X_test)\n", + "# Use version-agnostic RMSE calc (compatible with older sklearn)\n", + "test_rmse_C = np.sqrt(mean_squared_error(y_test, y_pred_C))\n", + "test_mae_C = mean_absolute_error(y_test, y_pred_C)\n", + "\n", + "print(\"=== Pipeline C: preproc1 + GradientBoostingRegressor (Advanced) ===\")\n", + "print(\"Best CV RMSE:\", round(best_cv_rmse_C, 4))\n", + "print(\"Test RMSE :\", round(test_rmse_C, 4))\n", + "print(\"Test MAE :\", round(test_mae_C, 4))\n", + "print(\"Best Params:\", gridC.best_params_)\n", + "\n", + "# Keep for later comparison/pickling/SHAP\n", + "best_pipe_C = gridC.best_estimator_\n" ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 23, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "c:\\Users\\bprabaharan\\Downloads\\Production\\production-env\\Lib\\site-packages\\sklearn\\preprocessing\\_data.py:2846: UserWarning: n_quantiles (517) is greater than the total number of samples (413). n_quantiles is set to n_samples.\n", + " warnings.warn(\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "=== Pipeline D: preproc2 + RandomForestRegressor (Advanced) ===\n", + "Best CV RMSE: 39.8771\n", + "Test RMSE : 108.0806\n", + "Test MAE : 24.117\n", + "Best Params: {'regressor__max_depth': 6, 'regressor__max_features': 'log2', 'regressor__min_samples_leaf': 4, 'regressor__n_estimators': 800}\n" + ] + } + ], "source": [ - "# Pipeline D = preproc2 + advanced model\n", + "# === Pipeline D: preproc2 + RandomForestRegressor (Advanced) ===\n", + "\n", + "from sklearn.pipeline import Pipeline\n", + "from sklearn.ensemble import RandomForestRegressor\n", + "from sklearn.model_selection import GridSearchCV\n", + "from sklearn.metrics import mean_squared_error, mean_absolute_error\n", + "import numpy as np\n", + "\n", + "RANDOM_STATE = 42\n", + "SCORING = \"neg_root_mean_squared_error\" # RMSE (negated for sklearn scoring)\n", + "CV = 5\n", + "N_JOBS = -1\n", + "\n", + "# Build Pipeline D using the existing preproc2\n", + "pipeD = Pipeline(steps=[\n", + " (\"preprocessing\", preproc2), # <-- assumes preproc2 already exists\n", + " (\"regressor\", RandomForestRegressor(\n", + " random_state=RANDOM_STATE,\n", + " n_jobs=N_JOBS\n", + " ))\n", + "])\n", + "\n", + "# Diverse, reasonable hyperparameter grid (>= 4 combos)\n", + "param_grid_D = {\n", + " \"regressor__n_estimators\": [200, 500, 800],\n", + " \"regressor__max_depth\": [None, 6, 12, 20],\n", + " \"regressor__min_samples_leaf\": [1, 2, 4],\n", + " \"regressor__max_features\": [\"sqrt\", \"log2\", None],\n", + "}\n", + "\n", + "gridD = GridSearchCV(\n", + " estimator=pipeD,\n", + " param_grid=param_grid_D,\n", + " scoring=SCORING,\n", + " cv=CV,\n", + " n_jobs=N_JOBS,\n", + " refit=True,\n", + " verbose=0,\n", + " return_train_score=False\n", + ")\n", + "\n", + "# Fit on training data\n", + "gridD.fit(X_train, y_train)\n", "\n", - " " + "# Evaluation\n", + "best_cv_rmse_D = -gridD.best_score_ # convert neg RMSE -> RMSE\n", + "y_pred_D = gridD.best_estimator_.predict(X_test)\n", + "\n", + "# Version-agnostic RMSE (works on older sklearn)\n", + "test_rmse_D = np.sqrt(mean_squared_error(y_test, y_pred_D))\n", + "test_mae_D = mean_absolute_error(y_test, y_pred_D)\n", + "\n", + "print(\"=== Pipeline D: preproc2 + RandomForestRegressor (Advanced) ===\")\n", + "print(\"Best CV RMSE:\", round(best_cv_rmse_D, 4))\n", + "print(\"Test RMSE :\", round(test_rmse_D, 4))\n", + "print(\"Test MAE :\", round(test_mae_D, 4))\n", + "print(\"Best Params:\", gridD.best_params_)\n", + "\n", + "# Keep for later comparison/pickling/SHAP\n", + "best_pipe_D = gridD.best_estimator_\n", + "\n" ] }, { @@ -276,38 +836,385 @@ }, { "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, + "execution_count": 24, "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "c:\\Users\\bprabaharan\\Downloads\\Production\\production-env\\Lib\\site-packages\\sklearn\\preprocessing\\_data.py:2846: UserWarning: n_quantiles (517) is greater than the total number of samples (413). n_quantiles is set to n_samples.\n", + " warnings.warn(\n", + "c:\\Users\\bprabaharan\\Downloads\\Production\\production-env\\Lib\\site-packages\\sklearn\\preprocessing\\_data.py:2846: UserWarning: n_quantiles (517) is greater than the total number of samples (413). n_quantiles is set to n_samples.\n", + " warnings.warn(\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "=== Grid Search Results (sorted by Test RMSE) ===\n", + " model best_cv_rmse test_rmse test_mae best_params\n", + " Pipeline D: preproc2 + RF 6.314834 108.080555 24.116965 {'regressor__max_depth': 6, 'regressor__max_features': 'log2', 'regressor__min_samples_leaf': 4, 'regressor__n_estimators': 800}\n", + "Pipeline A: preproc1 + Ridge 6.278958 108.351406 24.177821 {'regressor__alpha': 100.0, 'regressor__fit_intercept': True}\n", + "Pipeline B: preproc2 + Lasso 6.227727 108.636085 24.378057 {'regressor__alpha': 1.0, 'regressor__fit_intercept': True}\n", + " Pipeline C: preproc1 + GBR 6.723957 109.359080 25.206081 {'regressor__learning_rate': 0.01, 'regressor__max_depth': 2, 'regressor__n_estimators': 100, 'regressor__subsample': 0.8}\n", + "\n", + "Best overall model: Pipeline D: preproc2 + RF\n" + ] + } + ], + "source": [ + "# -------------------------------\n", + "# Run GridSearchCV for each\n", + "# -------------------------------\n", + "search_specs = [\n", + " (\"Pipeline A: preproc1 + Ridge\", pipeA, param_grid_A),\n", + " (\"Pipeline B: preproc2 + Lasso\", pipeB, param_grid_B),\n", + " (\"Pipeline C: preproc1 + GBR\", pipeC, param_grid_C),\n", + " (\"Pipeline D: preproc2 + RF\", pipeD, param_grid_D),\n", + "]\n", + "\n", + "results = []\n", + "all_searches = {}\n", + "\n", + "for name, pipe, grid in search_specs:\n", + " gs = GridSearchCV(\n", + " estimator=pipe,\n", + " param_grid=grid,\n", + " scoring=SCORING, # neg MSE\n", + " cv=CV,\n", + " n_jobs=N_JOBS,\n", + " refit=True,\n", + " verbose=0,\n", + " return_train_score=False\n", + " )\n", + " gs.fit(X_train, y_train)\n", + " all_searches[name] = gs\n", + "\n", + " # Convert cross-val neg MSE -> RMSE\n", + " best_cv_rmse = np.sqrt(-gs.best_score_)\n", + "\n", + " # Hold-out test evaluation\n", + " y_pred = gs.best_estimator_.predict(X_test)\n", + " test_rmse = np.sqrt(mean_squared_error(y_test, y_pred))\n", + " test_mae = mean_absolute_error(y_test, y_pred)\n", + "\n", + " results.append({\n", + " \"model\": name,\n", + " \"best_cv_rmse\": best_cv_rmse,\n", + " \"test_rmse\": test_rmse,\n", + " \"test_mae\": test_mae,\n", + " \"best_params\": gs.best_params_,\n", + " \"best_estimator\": gs.best_estimator_,\n", + " })\n", + "\n", + "# -------------------------------\n", + "# Summarize results\n", + "# -------------------------------\n", + "results_df = pd.DataFrame(results).sort_values(by=\"test_rmse\", ascending=True).reset_index(drop=True)\n", + "\n", + "# Display concise view (omit the estimator object for printing)\n", + "display_cols = [\"model\", \"best_cv_rmse\", \"test_rmse\", \"test_mae\", \"best_params\"]\n", + "print(\"=== Grid Search Results (sorted by Test RMSE) ===\")\n", + "print(results_df[display_cols].to_string(index=False))\n", + "\n", + "# Keep convenient handles for later steps (pickling, SHAP)\n", + "best_overall_idx = results_df[\"test_rmse\"].idxmin()\n", + "best_model_name = results_df.loc[best_overall_idx, \"model\"]\n", + "best_pipeline = results_df.loc[best_overall_idx, \"best_estimator\"]\n", + "\n", + "print(\"\\nBest overall model:\", best_model_name)\n" + ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 25, "metadata": {}, - "outputs": [], - "source": [] + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "=== Grid Search Results (sorted by Test RMSE) ===\n", + " model tuned_hyperparameters best_cv_rmse test_rmse test_mae best_params\n", + "Pipeline A: preproc1 + Ridge [alpha] 46.098980 108.351406 24.177821 {'regressor__alpha': 100.0}\n", + " Pipeline C: preproc1 + GBR [learning_rate, n_estimators] 56.994485 108.407404 25.232833 {'regressor__learning_rate': 0.01, 'regressor__n_estimators': 100}\n", + " Pipeline D: preproc2 + RF [n_estimators] 52.355982 108.613331 25.804381 {'regressor__n_estimators': 800}\n", + "Pipeline B: preproc2 + Lasso [alpha] 45.705590 108.636085 24.378057 {'regressor__alpha': 1.0}\n", + "\n", + "Best overall model: Pipeline A: preproc1 + Ridge\n" + ] + } + ], + "source": [ + "\n", + "# === Tune ≥1 hyperparameter per pipeline (A–D) with 5-fold CV, using an error metric ===\n", + "\n", + "from sklearn.model_selection import GridSearchCV\n", + "from sklearn.metrics import mean_squared_error, mean_absolute_error\n", + "import numpy as np\n", + "import pandas as pd\n", + "import warnings\n", + "\n", + "warnings.filterwarnings(\"ignore\")\n", + "\n", + "CV = 5\n", + "N_JOBS = -1\n", + "\n", + "# Use neg MSE for maximum compatibility across sklearn versions; we will convert to RMSE\n", + "SCORING = \"neg_mean_squared_error\"\n", + "\n", + "# -------------------------------\n", + "# Define grids (≥ 1 HP each)\n", + "# -------------------------------\n", + "# Pipeline A (preproc1 + Ridge): tune alpha\n", + "param_grid_A = {\n", + " \"regressor__alpha\": [0.01, 0.1, 1.0, 10.0, 100.0]\n", + "}\n", + "\n", + "# Pipeline B (preproc2 + Lasso): tune alpha\n", + "param_grid_B = {\n", + " \"regressor__alpha\": [0.0001, 0.0005, 0.001, 0.01, 0.1, 1.0]\n", + "}\n", + "\n", + "# Pipeline C (preproc1 + GradientBoostingRegressor): tune learning_rate (and n_estimators for better coverage)\n", + "param_grid_C = {\n", + " \"regressor__learning_rate\": [0.01, 0.05, 0.1],\n", + " \"regressor__n_estimators\": [100, 300, 600]\n", + "}\n", + "\n", + "# Pipeline D (preproc2 + RandomForestRegressor): tune n_estimators\n", + "param_grid_D = {\n", + " \"regressor__n_estimators\": [200, 500, 800]\n", + "}\n", + "\n", + "# -------------------------------\n", + "# Run GridSearchCV for each pipeline\n", + "# -------------------------------\n", + "search_specs = [\n", + " (\"Pipeline A: preproc1 + Ridge\", pipeA, param_grid_A),\n", + " (\"Pipeline B: preproc2 + Lasso\", pipeB, param_grid_B),\n", + " (\"Pipeline C: preproc1 + GBR\", pipeC, param_grid_C),\n", + " (\"Pipeline D: preproc2 + RF\", pipeD, param_grid_D),\n", + "]\n", + "\n", + "results = []\n", + "all_searches = {}\n", + "\n", + "for name, pipe, grid in search_specs:\n", + " gs = GridSearchCV(\n", + " estimator=pipe,\n", + " param_grid=grid,\n", + " scoring=SCORING, # neg MSE (we'll convert to RMSE)\n", + " cv=CV,\n", + " n_jobs=N_JOBS,\n", + " refit=True,\n", + " verbose=0,\n", + " return_train_score=False\n", + " )\n", + " gs.fit(X_train, y_train)\n", + " all_searches[name] = gs\n", + "\n", + " # Convert cross-val neg MSE -> RMSE for readability\n", + " best_cv_rmse = np.sqrt(-gs.best_score_)\n", + "\n", + " # Hold-out test evaluation\n", + " y_pred = gs.best_estimator_.predict(X_test)\n", + " test_rmse = np.sqrt(mean_squared_error(y_test, y_pred))\n", + " test_mae = mean_absolute_error(y_test, y_pred)\n", + "\n", + " # Keep a tidy record\n", + " tuned_keys = [k.replace(\"regressor__\", \"\") for k in grid.keys()]\n", + " results.append({\n", + " \"model\": name,\n", + " \"tuned_hyperparameters\": tuned_keys,\n", + " \"best_cv_rmse\": best_cv_rmse,\n", + " \"test_rmse\": test_rmse,\n", + " \"test_mae\": test_mae,\n", + " \"best_params\": gs.best_params_,\n", + " \"best_estimator\": gs.best_estimator_\n", + " })\n", + "\n", + "# -------------------------------\n", + "# Summarize results\n", + "# -------------------------------\n", + "results_df = pd.DataFrame(results).sort_values(by=\"test_rmse\", ascending=True).reset_index(drop=True)\n", + "\n", + "display_cols = [\"model\", \"tuned_hyperparameters\", \"best_cv_rmse\", \"test_rmse\", \"test_mae\", \"best_params\"]\n", + "print(\"=== Grid Search Results (sorted by Test RMSE) ===\")\n", + "print(results_df[display_cols].to_string(index=False))\n", + "\n", + "# Keep the winning pipeline handy for SHAP and export\n", + "best_overall_idx = results_df[\"test_rmse\"].idxmin()\n", + "best_model_name = results_df.loc[best_overall_idx, \"model\"]\n", + "best_pipeline = results_df.loc[best_overall_idx, \"best_estimator\"]\n", + "\n", + "print(\"\\nBest overall model:\", best_model_name)\n" + ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 26, "metadata": {}, - "outputs": [], - "source": [] + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "=== Expanded Grid Search Results (≥ 4 combos each; sorted by Test RMSE) ===\n", + " model num_grid_combos best_cv_rmse test_rmse test_mae best_params\n", + "Pipeline A: preproc1 + Ridge 5 46.098980 108.351406 24.177821 {'regressor__alpha': 100.0}\n", + " Pipeline D: preproc2 + RF 9 52.254808 108.616196 25.595651 {'regressor__max_depth': 10, 'regressor__n_estimators': 800}\n", + "Pipeline B: preproc2 + Lasso 6 45.705590 108.636085 24.378057 {'regressor__alpha': 1.0}\n", + " Pipeline C: preproc1 + GBR 18 51.583055 109.671090 25.390907 {'regressor__learning_rate': 0.01, 'regressor__max_depth': 2, 'regressor__n_estimators': 100}\n", + "\n", + "Best overall model: Pipeline A: preproc1 + Ridge\n" + ] + } + ], + "source": [ + "\n", + "# ==== Expanded GridSearch on Pipelines A–D (≥ 4 combos each) ====\n", + "\n", + "from sklearn.pipeline import Pipeline\n", + "from sklearn.linear_model import Ridge, Lasso\n", + "from sklearn.ensemble import GradientBoostingRegressor, RandomForestRegressor\n", + "from sklearn.model_selection import GridSearchCV\n", + "from sklearn.metrics import mean_squared_error, mean_absolute_error\n", + "import numpy as np\n", + "import pandas as pd\n", + "import warnings\n", + "\n", + "warnings.filterwarnings(\"ignore\")\n", + "\n", + "RANDOM_STATE = 42\n", + "CV = 5\n", + "N_JOBS = -1\n", + "\n", + "# Use neg MSE for broad compatibility; convert to RMSE\n", + "SCORING = \"neg_mean_squared_error\"\n", + "\n", + "# -------------------------------\n", + "# A) preproc1 + Ridge (≥ 4 combos)\n", + "# -------------------------------\n", + "pipeA = Pipeline(steps=[\n", + " (\"preprocessing\", preproc1),\n", + " (\"regressor\", Ridge())\n", + "])\n", + "\n", + "param_grid_A = {\n", + " \"regressor__alpha\": [0.01, 0.1, 1.0, 10.0, 100.0], # 5 combos\n", + " # Optionally add another knob (kept simple):\n", + " # \"regressor__fit_intercept\": [True, False] # would double combinations\n", + "}\n", + "\n", + "# -------------------------------\n", + "# B) preproc2 + Lasso (≥ 4 combos)\n", + "# -------------------------------\n", + "pipeB = Pipeline(steps=[\n", + " (\"preprocessing\", preproc2),\n", + " (\"regressor\", Lasso(max_iter=10000))\n", + "])\n", + "\n", + "param_grid_B = {\n", + " \"regressor__alpha\": [0.0001, 0.0005, 0.001, 0.01, 0.1, 1.0] # 6 combos\n", + "}\n", + "\n", + "# -------------------------------\n", + "# C) preproc1 + GradientBoostingRegressor (≥ 4 combos)\n", + "# -------------------------------\n", + "pipeC = Pipeline(steps=[\n", + " (\"preprocessing\", preproc1),\n", + " (\"regressor\", GradientBoostingRegressor(random_state=RANDOM_STATE))\n", + "])\n", + "\n", + "param_grid_C = {\n", + " \"regressor__n_estimators\": [100, 300, 600], # 3\n", + " \"regressor__learning_rate\": [0.01, 0.05, 0.1],# 3\n", + " \"regressor__max_depth\": [2, 3], # 2 --> 3×3×2 = 18 combos\n", + " # (You can reintroduce 'subsample' if you want even more diversity)\n", + "}\n", + "\n", + "# -------------------------------\n", + "# D) preproc2 + RandomForestRegressor (≥ 4 combos)\n", + "# -------------------------------\n", + "pipeD = Pipeline(steps=[\n", + " (\"preprocessing\", preproc2),\n", + " (\"regressor\", RandomForestRegressor(random_state=RANDOM_STATE, n_jobs=N_JOBS))\n", + "])\n", + "\n", + "# Ensure ≥ 4 combos: two knobs with 3 values each → 9 combos\n", + "param_grid_D = {\n", + " \"regressor__n_estimators\": [200, 500, 800], # 3\n", + " \"regressor__max_depth\": [None, 10, 20], # 3 --> 9 combos\n", + " # You can further expand:\n", + " # \"regressor__min_samples_leaf\": [1, 2],\n", + " # \"regressor__max_features\": [\"sqrt\", \"log2\"]\n", + "}\n", + "\n", + "# -------------------------------\n", + "# Run GridSearchCV for each\n", + "# -------------------------------\n", + "search_specs = [\n", + " (\"Pipeline A: preproc1 + Ridge\", pipeA, param_grid_A),\n", + " (\"Pipeline B: preproc2 + Lasso\", pipeB, param_grid_B),\n", + " (\"Pipeline C: preproc1 + GBR\", pipeC, param_grid_C),\n", + " (\"Pipeline D: preproc2 + RF\", pipeD, param_grid_D),\n", + "]\n", + "\n", + "results = []\n", + "all_searches = {}\n", + "\n", + "for name, pipe, grid in search_specs:\n", + " gs = GridSearchCV(\n", + " estimator=pipe,\n", + " param_grid=grid,\n", + " scoring=SCORING, # neg MSE (convert later)\n", + " cv=CV,\n", + " n_jobs=N_JOBS,\n", + " refit=True,\n", + " verbose=0,\n", + " return_train_score=False\n", + " )\n", + " gs.fit(X_train, y_train)\n", + " all_searches[name] = gs\n", + "\n", + " # Convert cross-val neg MSE -> RMSE\n", + " best_cv_rmse = np.sqrt(-gs.best_score_)\n", + "\n", + " # Hold-out test evaluation\n", + " y_pred = gs.best_estimator_.predict(X_test)\n", + " test_rmse = np.sqrt(mean_squared_error(y_test, y_pred))\n", + " test_mae = mean_absolute_error(y_test, y_pred)\n", + "\n", + " # Record results\n", + " results.append({\n", + " \"model\": name,\n", + " \"num_grid_combos\": np.prod([len(v) for v in grid.values()]),\n", + " \"best_cv_rmse\": best_cv_rmse,\n", + " \"test_rmse\": test_rmse,\n", + " \"test_mae\": test_mae,\n", + " \"best_params\": gs.best_params_,\n", + " \"best_estimator\": gs.best_estimator_\n", + " })\n", + "\n", + "# -------------------------------\n", + "# Summarize results\n", + "# -------------------------------\n", + "results_df = pd.DataFrame(results).sort_values(by=\"test_rmse\", ascending=True).reset_index(drop=True)\n", + "\n", + "display_cols = [\"model\", \"num_grid_combos\", \"best_cv_rmse\", \"test_rmse\", \"test_mae\", \"best_params\"]\n", + "print(\"=== Expanded Grid Search Results (≥ 4 combos each; sorted by Test RMSE) ===\")\n", + "print(results_df[display_cols].to_string(index=False))\n", + "\n", + "# Keep the winning pipeline handy for SHAP and export\n", + "best_overall_idx = results_df[\"test_rmse\"].idxmin()\n", + "best_model_name = results_df.loc[best_overall_idx, \"model\"]\n", + "best_pipeline = results_df.loc[best_overall_idx, \"best_estimator\"]\n", + "\n", + "print(\"\\nBest overall model:\", best_model_name)\n" + ] }, { "cell_type": "markdown", @@ -315,7 +1222,9 @@ "source": [ "# Evaluate\n", "\n", - "+ Which model has the best performance?" + "+ Which model has the best performance?\n", + "\n", + "Pipeline A wins on the core objective: lowest generalization error.\n" ] }, { @@ -329,17 +1238,36 @@ }, { "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, + "execution_count": 27, "metadata": {}, - "outputs": [], - "source": [] + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Saved best pipeline to: C:\\Users\\bprabaharan\\Downloads\\production\\02_activities\\artifacts\\forestfires_best_pipeline.pkl\n" + ] + } + ], + "source": [ + "\n", + "import pickle\n", + "from pathlib import Path\n", + "best_pipeline # -> fitted Pipeline object of the winner\n", + "\n", + "# Create an artifacts directory (relative to your notebook)\n", + "artifacts_dir = Path(\"../artifacts\")\n", + "artifacts_dir.mkdir(parents=True, exist_ok=True)\n", + "\n", + "# Choose a filename (include model name and a brief tag)\n", + "best_model_path = artifacts_dir / \"forestfires_best_pipeline.pkl\"\n", + "\n", + "# Save with a context manager\n", + "with open(best_model_path, \"wb\") as f:\n", + " pickle.dump(best_pipeline, f)\n", + "\n", + "print(f\"Saved best pipeline to: {best_model_path.resolve()}\")\n" + ] }, { "cell_type": "markdown", @@ -360,21 +1288,197 @@ "cell_type": "code", "execution_count": null, "metadata": {}, - "outputs": [], - "source": [] + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "\n", + "import shap\n", + "import numpy as np\n", + "import pandas as pd\n", + "\n", + "# --- Use best Pipeline A: preproc1 + Ridge ---\n", + "\n", + "best_pipeline # already created earlier\n", + "\n", + "\n", + "# Split the pipeline into preprocessing and model\n", + "preproc = best_pipeline.named_steps[\"preprocessing\"]\n", + "model = best_pipeline.named_steps[\"regressor\"] # Ridge\n", + "\n", + "# Transform data into the model's feature space\n", + "X_train_trans = preproc.transform(X_train)\n", + "X_test_trans = preproc.transform(X_test)\n", + "\n", + "# Get the post-transform feature names (OHE-expanded)\n", + "try:\n", + " feature_names = preproc.get_feature_names_out()\n", + "except Exception:\n", + " feature_names = np.array([f\"f{i}\" for i in range(X_train_trans.shape[1])])\n", + "\n", + "# Explain with LinearExplainer (ideal for linear models)\n", + "explainer = shap.LinearExplainer(model, X_train_trans, feature_perturbation=\"interventional\")\n", + "\n", + "# Call the explainer - returns an Explanation object\n", + "sv_train = explainer(X_train_trans) # global explanations on training data\n", + "sv_test = explainer(X_test_trans) # explanations on test data for local plots\n", + "\n", + "# Attach feature names for some plot helpers\n", + "try:\n", + " sv_train.feature_names = feature_names\n", + " sv_test.feature_names = feature_names\n", + "except Exception:\n", + " pass\n", + "\n", + "# --- GLOBAL explanations ---\n", + "# Beeswarm (distribution of SHAP impacts across training samples)\n", + "shap.summary_plot(sv_train.values, X_train_trans, feature_names=feature_names, plot_type=\"dot\", show=True)\n", + "\n", + "# Bar chart (mean |SHAP| = global importance ranking)\n", + "shap.summary_plot(sv_train.values, X_train_trans, feature_names=feature_names, plot_type=\"bar\", show=True)\n", + "\n", + "# --- LOCAL explanation for a single test observation ---\n", + "i = 0 # pick any valid index from the test set\n", + "shap.plots.waterfall(sv_test[i], max_display=15) # modern API expects an Explanation row\n" + ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 37, "metadata": {}, - "outputs": [], - "source": [] + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Top 10 features:\n", + " feature mean_abs_shap\n", + "0 num__dmc 3.369488\n", + "1 num__coord_x 3.145918\n", + "2 num__temp 1.722197\n", + "3 num__rh 1.505499\n", + "4 num__wind 1.047723\n", + "5 num__dc 0.838974\n", + "6 cat__day_fri 0.568845\n", + "7 num__coord_y 0.558082\n", + "8 num__isi 0.438313\n", + "9 cat__day_thu 0.381790\n", + "\n", + "Bottom 10 features (candidates to remove):\n", + " feature mean_abs_shap\n", + "19 num__ffmc 0.030183\n", + "20 cat__month_aug 0.024302\n", + "21 cat__day_mon 0.022402\n", + "22 cat__month_dec 0.020541\n", + "23 cat__month_apr 0.012349\n", + "24 cat__day_wed 0.005597\n", + "25 cat__month_oct 0.004866\n", + "26 cat__month_jan 0.001011\n", + "27 cat__month_may 0.000818\n", + "28 cat__month_nov 0.000174\n", + "Baseline: RMSE=108.3514, MAE=24.1778\n", + "Pruned : RMSE=108.3514, MAE=24.1778\n", + "Best params (pruned): {'regressor__alpha': 100.0}\n" + ] + } + ], + "source": [ + "#DROP LOW-IMPORTANCE FEATURES BASED ON SHAP VALUES\n", + "mean_abs_shap = np.abs(sv_train.values).mean(axis=0)\n", + "global_importance = (\n", + " pd.DataFrame({\"feature\": feature_names, \"mean_abs_shap\": mean_abs_shap})\n", + " .sort_values(\"mean_abs_shap\", ascending=False)\n", + " .reset_index(drop=True)\n", + ")\n", + "\n", + "print(\"Top 10 features:\\n\", global_importance.head(10))\n", + "print(\"\\nBottom 10 features (candidates to remove):\\n\", global_importance.tail(10))\n", + "\n", + "# Drop 10 low-importance raw columns\n", + "cols_to_drop = [\"cat__month_may\", \"cat__month_nov\", \"num__ffmc\", \"cat__month_aug\", \"cat__day_mon\", \"cat__month_dec\", \"cat__month_apr\", \"cat__day_wed\", \"cat__month_oct\", \"cat__month_jan\"] \n", + "\n", + "feature_cols_pruned = [c for c in X.columns if c not in cols_to_drop]\n", + "num_pruned = [c for c in feature_cols_pruned if c in X.select_dtypes(include=np.number).columns]\n", + "cat_pruned = [c for c in feature_cols_pruned if c in X.select_dtypes(include=[\"object\",\"category\"]).columns]\n", + "\n", + "from sklearn.compose import ColumnTransformer\n", + "from sklearn.preprocessing import OneHotEncoder, StandardScaler\n", + "from sklearn.pipeline import Pipeline\n", + "from sklearn.model_selection import GridSearchCV\n", + "\n", + "preproc1_pruned = ColumnTransformer(\n", + " [(\"num\", Pipeline([(\"scaler\", StandardScaler())]), num_pruned),\n", + " (\"cat\", Pipeline([(\"ohe\", OneHotEncoder(handle_unknown=\"ignore\", sparse_output=False))]), cat_pruned)],\n", + " remainder=\"drop\",\n", + " verbose_feature_names_out=True\n", + ")\n", + "\n", + "pipeA_pruned = Pipeline([\n", + " (\"preprocessing\", preproc1_pruned),\n", + " (\"regressor\", model.__class__()) # Ridge()\n", + "])\n", + "\n", + "gridA = GridSearchCV(\n", + " pipeA_pruned,\n", + " {\"regressor__alpha\": [0.01, 0.1, 1.0, 10.0, 100.0]},\n", + " scoring=\"neg_mean_squared_error\",\n", + " cv=5, n_jobs=-1, refit=True\n", + ")\n", + "gridA.fit(X_train[feature_cols_pruned], y_train)\n", + "\n", + "# Compare hold-out performance\n", + "y_pred_base = best_pipeline.predict(X_test)\n", + "y_pred_prune = gridA.best_estimator_.predict(X_test[feature_cols_pruned])\n", + "\n", + "rmse_base = np.sqrt(((y_test - y_pred_base )**2).mean()); mae_base = np.abs(y_test - y_pred_base).mean()\n", + "rmse_prune = np.sqrt(((y_test - y_pred_prune)**2).mean()); mae_prune = np.abs(y_test - y_pred_prune).mean()\n", + "\n", + "print(f\"Baseline: RMSE={rmse_base:.4f}, MAE={mae_base:.4f}\")\n", + "print(f\"Pruned : RMSE={rmse_prune:.4f}, MAE={mae_prune:.4f}\")\n", + "print(\"Best params (pruned):\", gridA.best_params_)\n" + ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "*(Answer here.)*" + "*(Answer here.)*\n", + "We use SHAP to calculate how much each feature contributes to moving the prediction from the average value. The higher the SHAP value the more it contributes. The closer to 0 it is the less it contributes (and can be considered for removal)\n", + "\n", + "We test this by updating the pipeline to drop these features and comparing the difference in the RMSE & MAE values. If the RMSE/MAE values stay the same or become lower then we know that the features added little predicitve value. If the values got higher then the features should stay in the model.\n", + "\n", + "Dropping the 10 lowest importance features had no change on the RMSE & MAE as they contributed very little to the prediction. Ridge was already neutralizing the weak variables. Results confirm the SHAP analysis was correct and model is robust to irrelvant features. \n", + "\n", + "\n" ] }, { @@ -423,7 +1527,7 @@ ], "metadata": { "kernelspec": { - "display_name": "env", + "display_name": "production-env", "language": "python", "name": "python3" }, @@ -437,7 +1541,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.12.3" + "version": "3.11.14" } }, "nbformat": 4,