5.Model review.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.metrics import (\n",
    "    r2_score, mean_squared_error, mean_absolute_error,\n",
    "    accuracy_score, precision_score, recall_score, f1_score,\n",
    "    confusion_matrix, roc_curve, auc\n",
    ")\n",
    "from sklearn.model_selection import cross_val_score, KFold\n",
    "from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor\n",
    "from sklearn.datasets import make_classification, make_regression\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# 回帰問題の評価指標\n",
    "def evaluate_regression(y_true, y_pred):\n",
    "    \"\"\"\n",
    "    回帰モデルの評価指標を計算する関数\n",
    "    \"\"\"\n",
    "    r2 = r2_score(y_true, y_pred)\n",
    "    rmse = np.sqrt(mean_squared_error(y_true, y_pred))\n",
    "    mae = mean_absolute_error(y_true, y_pred)\n",
    "    return {\"R^2\": r2, \"RMSE\": rmse, \"MAE\": mae}\n",
    "\n",
    "# 分類問題の評価指標\n",
    "def evaluate_classification(y_true, y_pred):\n",
    "    \"\"\"\n",
    "    分類モデルの評価指標を計算する関数\n",
    "    \"\"\"\n",
    "    accuracy = accuracy_score(y_true, y_pred)\n",
    "    precision = precision_score(y_true, y_pred, average='weighted')\n",
    "    recall = recall_score(y_true, y_pred, average='weighted')\n",
    "    f1 = f1_score(y_true, y_pred, average='weighted')\n",
    "    return {\"Accuracy\": accuracy, \"Precision\": precision, \"Recall\": recall, \"F1 Score\": f1}\n",
    "\n",
    "# 交差検証\n",
    "def cross_validation(model, X, y, k=5):\n",
    "    \"\"\"\n",
    "    K-Fold 交差検証を実行する関数\n",
    "    \"\"\"\n",
    "    kf = KFold(n_splits=k, shuffle=True, random_state=42)\n",
    "    scores = cross_val_score(model, X, y, cv=kf)\n",
    "    return scores.mean()\n",
    "\n",
    "# 混同行列の可視化\n",
    "def plot_confusion_matrix(y_true, y_pred):\n",
    "    \"\"\"\n",
    "    混同行列をプロットする関数\n",
    "    \"\"\"\n",
    "    cm = confusion_matrix(y_true, y_pred)\n",
    "    plt.figure(figsize=(6, 4))\n",
    "    sns.heatmap(cm, annot=True, fmt='d', cmap='Blues', xticklabels=[\"Negative\", \"Positive\"], yticklabels=[\"Negative\", \"Positive\"])\n",
    "    plt.xlabel(\"Predicted\")\n",
    "    plt.ylabel(\"Actual\")\n",
    "    plt.title(\"Confusion Matrix\")\n",
    "    plt.show()\n",
    "\n",
    "# ROC曲線とAUC\n",
    "def plot_roc_curve(y_true, y_scores):\n",
    "    \"\"\"\n",
    "    ROC曲線とAUCをプロットする関数\n",
    "    \"\"\"\n",
    "    fpr, tpr, _ = roc_curve(y_true, y_scores)\n",
    "    roc_auc = auc(fpr, tpr)\n",
    "    plt.figure(figsize=(6, 4))\n",
    "    plt.plot(fpr, tpr, color='blue', label=f'AUC = {roc_auc:.2f}')\n",
    "    plt.plot([0, 1], [0, 1], linestyle='--', color='gray')\n",
    "    plt.xlabel(\"False Positive Rate\")\n",
    "    plt.ylabel(\"True Positive Rate\")\n",
    "    plt.title(\"ROC Curve\")\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "    return roc_auc\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "回帰評価指標: {'R^2': 0.9993636078552096, 'RMSE': 0.9419524252704788, 'MAE': 0.6034531651339912}\n",
      "分類評価指標: {'Accuracy': 0.9, 'Precision': 0.9222222222222222, 'Recall': 0.9, 'F1 Score': 0.9020833333333333}\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC AUC: 1.0\n"
     ]
    }
   ],
   "source": [
    "## 回帰問題\n",
    "X_reg, y_reg = make_regression(n_samples=100, n_features=1, noise=0.1, random_state=42)\n",
    "X_train_reg, X_test_reg, y_train_reg, y_test_reg = train_test_split(X_reg, y_reg, test_size=0.2, random_state=42)\n",
    "model_reg = RandomForestRegressor(n_estimators=100, random_state=42)\n",
    "model_reg.fit(X_train_reg, y_train_reg)\n",
    "y_pred_reg = model_reg.predict(X_test_reg)\n",
    "regression_metrics = evaluate_regression(y_test_reg, y_pred_reg)\n",
    "print(\"回帰評価指標:\", regression_metrics)\n",
    "\n",
    "## 分類問題\n",
    "X_clf, y_clf = make_classification(n_samples=100, n_features=5, random_state=42)\n",
    "X_train_clf, X_test_clf, y_train_clf, y_test_clf = train_test_split(X_clf, y_clf, test_size=0.2, random_state=42)\n",
    "model_clf = RandomForestClassifier(n_estimators=100, random_state=42)\n",
    "model_clf.fit(X_train_clf, y_train_clf)\n",
    "y_pred_clf = model_clf.predict(X_test_clf)\n",
    "classification_metrics = evaluate_classification(y_test_clf, y_pred_clf)\n",
    "print(\"分類評価指標:\", classification_metrics)\n",
    "\n",
    "# 混同行列のプロット\n",
    "plot_confusion_matrix(y_test_clf, y_pred_clf)\n",
    "\n",
    "# ROC曲線とAUCのプロット\n",
    "y_scores_clf = model_clf.predict_proba(X_test_clf)[:, 1]\n",
    "roc_auc_value = plot_roc_curve(y_test_clf, y_scores_clf)\n",
    "print(\"ROC AUC:\", roc_auc_value)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "回帰モデルの交差検証スコア（平均 R^2）： 0.7297421575998921\n"
     ]
    }
   ],
   "source": [
    "# データの生成（回帰問題）\n",
    "X_reg, y_reg = make_regression(n_samples=100, n_features=5, noise=0.1, random_state=42)\n",
    "\n",
    "# モデルの定義（ランダムフォレスト回帰）\n",
    "model_reg = RandomForestRegressor(n_estimators=100, random_state=42)\n",
    "\n",
    "# 交差検証の実行\n",
    "reg_cv_score = cross_validation(model_reg, X_reg, y_reg, k=5)\n",
    "\n",
    "print(\"回帰モデルの交差検証スコア（平均 R^2）：\", reg_cv_score)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "分類モデルの交差検証スコア（平均 Accuracy）： 0.9200000000000002\n"
     ]
    }
   ],
   "source": [
    "# データの生成（分類問題）\n",
    "X_clf, y_clf = make_classification(n_samples=100, n_features=5, random_state=42)\n",
    "\n",
    "# モデルの定義（ランダムフォレスト分類器）\n",
    "model_clf = RandomForestClassifier(n_estimators=100, random_state=42)\n",
    "\n",
    "# 交差検証の実行\n",
    "clf_cv_score = cross_validation(model_clf, X_clf, y_clf, k=5)\n",
    "\n",
    "print(\"分類モデルの交差検証スコア（平均 Accuracy）：\", clf_cv_score)\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "dataSci",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.20"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}