noshot 8.0.0__py3-none-any.whl → 10.0.0__py3-none-any.whl

noshot/data/ML TS XAI/ML/ML Lab H Sec/3. Insurance Target Categorical (Overfit vs Regularized).ipynb

@@ -0,0 +1,274 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "55832f35-459a-4a0e-a379-35484f92d1c5",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import tensorflow as tf\n",
+    "from tensorflow.keras.models import Sequential\n",
+    "from tensorflow.keras.layers import Dense, Dropout\n",
+    "from tensorflow.keras.optimizers import Adam\n",
+    "from tensorflow.keras import regularizers\n",
+    "from sklearn.preprocessing import LabelEncoder\n",
+    "from sklearn.model_selection import train_test_split\n",
+    "from sklearn.metrics import classification_report, confusion_matrix \n",
+    "from sklearn.metrics import ConfusionMatrixDisplay, accuracy_score, roc_curve, auc\n",
+    "\n",
+    "import warnings\n",
+    "warnings.filterwarnings('ignore')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f3344d11-95e5-481b-93e4-318bf834d2c0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df = pd.read_csv(\"insurance.csv\")\n",
+    "df.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "62045487-f95d-402a-b18e-85009ff58903",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "labels = ['Low', 'Medium', 'High', 'Very High']\n",
+    "df['charges'] = pd.qcut(df['charges'], q=4, labels=labels)\n",
+    "df.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "eb3bd408-f8a3-41fd-a778-e7608071e5e4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "cols = ['sex', 'smoker', 'region', 'charges']\n",
+    "df[cols] = df[cols].apply(LabelEncoder().fit_transform)\n",
+    "df.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c07602a7-b504-4d9b-a3c6-68bceee8280b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "X = df.drop(columns=['charges'])\n",
+    "y = tf.keras.utils.to_categorical(df['charges'], num_classes=4)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "89d7dd41-4d03-4f50-8639-9065bd17b7f6",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "overfit = Sequential([\n",
+    "    Dense(64, activation='relu', input_shape=(6,)),\n",
+    "    Dense(64, activation='relu'),\n",
+    "    Dense(4, activation='softmax')\n",
+    "])\n",
+    "\n",
+    "overfit.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8d55f010-fe99-40a4-a7f7-8fbe26e15e71",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "x_train, x_test, y_train, y_test = train_test_split(X, y, test_size=0.4, random_state=0)\n",
+    "history1 = overfit.fit(x_train, y_train, epochs=25, \n",
+    "                    batch_size=16, validation_split=0.2,\n",
+    "                    verbose=1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1151c540-a049-4cde-b6ec-51e214cf52ba",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "regularized = Sequential([\n",
+    "    Dense(128, activation='relu', input_shape=(6,), kernel_regularizer=regularizers.l2(0.001)),\n",
+    "    Dropout(0.1),\n",
+    "    Dense(64, activation='relu', kernel_regularizer=regularizers.l2(0.001)),\n",
+    "    Dropout(0.1),\n",
+    "    Dense(4, activation='softmax')\n",
+    "])\n",
+    "\n",
+    "regularized.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "88058b99-e8cd-47b6-a4bb-bc626996deca",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "early_stopping = tf.keras.callbacks.EarlyStopping(monitor='val_loss', \n",
+    "                                                  patience=10, \n",
+    "                                                  restore_best_weights=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "22964de8-9838-4f7c-9141-bffdf4def8b1",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "x_train, x_test, y_train, y_test = train_test_split(X, y, test_size=0.1, random_state=1)\n",
+    "history2 = regularized.fit(x_train, y_train, epochs=25, \n",
+    "                    batch_size=16, validation_split=0.3,\n",
+    "                    callbacks=[early_stopping], verbose=1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "829e7156-0ce9-4e48-becc-48d8c607fc4d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def plot_history(history1, history2):\n",
+    "    plt.figure(figsize=(12, 5))\n",
+    "    plt.subplot(1, 2, 1)\n",
+    "    plt.plot(history1.history['accuracy'], label='Overfit Model Validation Accuracy')\n",
+    "    plt.plot(history2.history['accuracy'], label='Regularized Model Validation Accuracy')\n",
+    "    plt.title('Training Accuracy')\n",
+    "    plt.xlabel('Epoch')\n",
+    "    plt.ylabel('Accuracy')\n",
+    "    plt.legend()\n",
+    "\n",
+    "    plt.subplot(1, 2, 2)\n",
+    "    plt.plot(history1.history['val_accuracy'], label='Overfit Model Validation Accuracy')\n",
+    "    plt.plot(history2.history['val_accuracy'], label='Regularized Model Validation Accuracy')\n",
+    "    plt.title('Validation Accuracy')\n",
+    "    plt.xlabel('Epoch')\n",
+    "    plt.ylabel('Accuracy')\n",
+    "    plt.legend()\n",
+    "\n",
+    "    plt.tight_layout()\n",
+    "    plt.show()\n",
+    "\n",
+    "plot_history(history1, history2)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "6ff519cf-eca6-48fe-923c-aaa18721608e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "loss1, accuracy1 = overfit.evaluate(x_train, y_train)\n",
+    "loss2, accuracy2 = overfit.evaluate(x_test, y_test)\n",
+    "print(f'Train accuracy Overfit Model: {accuracy1:.4f}')\n",
+    "print(f'Test accuracy Overfit Model: {accuracy2:.4f}')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "0aa14e69-14d8-4566-a532-59d17b89dd04",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "loss1, accuracy1 = regularized.evaluate(x_train, y_train)\n",
+    "loss2, accuracy2 = regularized.evaluate(x_test, y_test)\n",
+    "print(f'Train accuracy Regularized Model: {accuracy1:.4f}')\n",
+    "print(f'Test accuracy Regularized Model: {accuracy2:.4f}')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d1aadc08-5ef6-4285-b889-d4dad9ddb3fd",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "y_pred1 = np.argmax(overfit.predict(x_test), axis=1)\n",
+    "y_test1 = np.argmax(y_test, axis=1)\n",
+    "y_pred2 = np.argmax(regularized.predict(x_test), axis=1)\n",
+    "y_test2 = np.argmax(y_test, axis=1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c36a9495-2a6b-47da-b50a-060d0166412c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(\"Classification Report for Overfit:\")\n",
+    "print(classification_report(y_test1, y_pred1))\n",
+    "\n",
+    "print(\"\\nClassification Report for Regularized:\")\n",
+    "print(classification_report(y_test2, y_pred2))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8dec9763-db18-4476-9f2a-acfc3bd5b63b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "labels = df['charges'].unique()\n",
+    "cm1 = confusion_matrix(y_test1, y_pred1)\n",
+    "cm2 = confusion_matrix(y_test2, y_pred2)\n",
+    "\n",
+    "fig, ax = plt.subplots(1, 2, figsize=(14, 6))\n",
+    "\n",
+    "ConfusionMatrixDisplay(cm1, display_labels=labels).plot(ax=ax[0])\n",
+    "ax[0].set_title(\"Overfit Confusion Matrix\", size=25)\n",
+    "\n",
+    "ConfusionMatrixDisplay(cm2, display_labels=labels).plot(ax=ax[1])\n",
+    "ax[1].set_title(\"Regularized Confusion Matrix\", size=25)\n",
+    "\n",
+    "plt.tight_layout()\n",
+    "plt.show()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "NEW-VENV-1",
+   "language": "python",
+   "name": "new-venv-1"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.5"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

noshot/data/ML TS XAI/ML/ML Lab H Sec/3. Insurance Target Numerical (Overfit vs Regularized).ipynb

@@ -0,0 +1,263 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "bd31d202-9ef3-419d-85aa-478e28199448",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import tensorflow as tf\n",
+    "from tensorflow.keras.layers import Input, Dense, Dropout\n",
+    "from tensorflow.keras.models import Sequential\n",
+    "from tensorflow.keras.regularizers import l2\n",
+    "from tensorflow.keras.callbacks import EarlyStopping\n",
+    "from sklearn.preprocessing import StandardScaler, LabelEncoder\n",
+    "from sklearn.model_selection import train_test_split\n",
+    "from sklearn.metrics import mean_squared_error, r2_score"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "10fef1a2-2c53-42b4-b8ec-f1dae5b14a1c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df = pd.read_csv('insurance.csv')\n",
+    "df.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "6079f0b7-b6b3-493e-8a73-ff3f944db314",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "cat_cols = ['sex', 'smoker', 'region']\n",
+    "df[cat_cols] = df[cat_cols].apply(LabelEncoder().fit_transform)\n",
+    "print(df.head())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4a4d03e5-a817-4b66-9ab6-d32b18fe441a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "X = df.drop(columns='charges')\n",
+    "y = df['charges']\n",
+    "X = StandardScaler().fit_transform(X)\n",
+    "X_train_full, X_test, y_train_full, y_test = train_test_split(X, y, test_size=0.2, \n",
+    "                                                              random_state=42)\n",
+    "X_train, X_val, y_train, y_val = train_test_split(X_train_full, y_train_full, test_size=0.2, \n",
+    "                                                  random_state=42)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f4c58e24-e103-4d92-993c-6db9f9b57657",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "overfit_model = Sequential([\n",
+    "    Input(shape=(X_train.shape[1],)),\n",
+    "    Dense(64, activation='relu'),\n",
+    "    Dense(64, activation='relu'),\n",
+    "    Dense(1, activation='linear'),\n",
+    "])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "533969a3-9b46-4fd7-a21a-968e96821c08",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "reg_model = Sequential([\n",
+    "    Input(shape=(X_train.shape[1],)),\n",
+    "    Dense(64, activation='relu'),\n",
+    "    Dropout(0.1),\n",
+    "    Dense(64, activation='relu', kernel_regularizer=l2(0.001)),\n",
+    "    Dropout(0.1),\n",
+    "    Dense(1, activation='linear'),\n",
+    "])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c9471a5d-3c49-4531-bae6-8138c9812261",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(\"\\nOverfit Model Summary:\")\n",
+    "overfit_model.summary()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3efc2e9b-2b1e-416c-888e-9ea092de9d3f",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(\"\\nRegularized Model Summary:\")\n",
+    "reg_model.summary()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c76d9120-15c2-447b-9e9c-83c16c118096",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "overfit_model.compile(optimizer='adam', loss='mean_squared_error', metrics=['mae', 'mse'])\n",
+    "reg_model.compile(optimizer='adam', loss='mean_squared_error', metrics=['mae', 'mse'])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c05813ae-dd34-4caf-b008-9e683216c891",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "history_overfit = overfit_model.fit(\n",
+    "    X_train, y_train,\n",
+    "    validation_data=(X_val, y_val),\n",
+    "    epochs=20,\n",
+    "    batch_size=16,\n",
+    "    verbose=1\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "124d821f-3272-4a0b-9a24-0fedc7f7cfee",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "history_reg = reg_model.fit(\n",
+    "    X_train, y_train,\n",
+    "    validation_data=(X_val, y_val),\n",
+    "    epochs=20,\n",
+    "    batch_size=16,\n",
+    "    verbose=1\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "856e0e03-f681-4061-83b9-83da2823adda",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.figure(figsize=(14, 6))\n",
+    "plt.subplot(1, 2, 1)\n",
+    "plt.plot(history_overfit.history['val_loss'], label='Overfit Val Loss', color='blue', linestyle='--')\n",
+    "plt.plot(history_reg.history['val_loss'], label='Regularized Val Loss', color='brown', linestyle='--')\n",
+    "plt.title('Model Loss Comparison (MSE)')\n",
+    "plt.xlabel('Epoch')\n",
+    "plt.ylabel('Mean Squared Error Loss')\n",
+    "plt.legend()\n",
+    "plt.grid(True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f5315d68-df04-4883-b021-5b3d1be6c180",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.plot(history_overfit.history['val_mae'], label='Overfit Val MAE', color='blue')\n",
+    "plt.plot(history_reg.history['val_mae'], label='Regularized Val MAE', color='brown')\n",
+    "plt.title('Model Mean Absolute Error Comparison')\n",
+    "plt.xlabel('Epoch')\n",
+    "plt.ylabel('Mean Absolute Error')\n",
+    "plt.legend()\n",
+    "plt.grid(True)\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "bef16d6a-c694-495e-91b8-4b6021146b21",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "loss_overfit, mae_overfit, mse_overfit_eval = overfit_model.evaluate(X_test, y_test, verbose=0)\n",
+    "y_pred_overfit = overfit_model.predict(X_test, verbose=0).flatten()\n",
+    "r2_overfit = r2_score(y_test, y_pred_overfit)\n",
+    "\n",
+    "loss_reg, mae_reg, mse_reg_eval = reg_model.evaluate(X_test, y_test, verbose=0)\n",
+    "y_pred_reg = reg_model.predict(X_test, verbose=0).flatten()\n",
+    "r2_reg = r2_score(y_test, y_pred_reg)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cce008a5-7588-4574-a2a9-5b34f12038d5",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(\"\\nTest Set Performance:\")\n",
+    "print(f\"Overfit Model     - MSE: {mse_overfit_eval:.2f}, MAE: {mae_overfit:.2f}, R2: {r2_overfit:.4f}\")\n",
+    "print(f\"Regularized Model - MSE: {mse_reg_eval:.2f}, MAE: {mae_reg:.2f}, R2: {r2_reg:.4f}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "40c9647b-55d7-4d9a-a0ae-061f71cbec72",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.figure(figsize=(10, 7))\n",
+    "plt.scatter(y_test, y_pred_overfit, color='skyblue', edgecolors='k', linewidth=0.5, label=f'Overfit Model (R2={r2_overfit:.3f})')\n",
+    "plt.scatter(y_test, y_pred_reg, color='sandybrown', edgecolors='k', linewidth=0.5, label=f'Regularized Model (R2={r2_reg:.3f})')\n",
+    "plt.plot([min(y_test), max(y_test)], [min(y_test), max(y_test)], color='black', linestyle='--', linewidth=2, label='Perfect Prediction')\n",
+    "plt.title('Actual vs Predicted Charges (Test Set)')\n",
+    "plt.xlabel('Actual Charges ($)')\n",
+    "plt.ylabel('Predicted Charges ($)')\n",
+    "plt.legend()\n",
+    "plt.grid(True)\n",
+    "plt.show()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "NEW-VENV-1",
+   "language": "python",
+   "name": "new-venv-1"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.5"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}