from sklearn.linear_model import LogisticRegression from sklearn.neighbors import KNeighborsClassifier from sklearn.svm import SVC # モデルのリスト models = { ‘DecisionTree’: DecisionTreeClassifier(max_depth=3, random_state=42), ‘LogisticRegression’: LogisticRegression(max_iter=200, random_state=42), ‘KNN’: KNeighborsClassifier(n_neighbors=5), ‘SVM’: SVC(kernel=’rbf’, random_state=42) } # 各モデルを訓練・評価 results = {} for name, model in models.items(): # 訓練 model.fit(X_train, y_train) # 評価 train_score = model.score(X_train, y_train) test_score = model.score(X_test, y_test) results[name] = { ‘train’: train_score, ‘test’: test_score } print(f”{name:20s} – Train: {train_score:.4f}, Test: {test_score:.4f}”) # 結果を可視化 import pandas as pd df_results = pd.DataFrame(results).T df_results.plot(kind=’bar’, figsize=(10, 6)) plt.title(‘Model Comparison’) plt.ylabel(‘Accuracy’) plt.xlabel(‘Model’) plt.ylim(0.9, 1.01) plt.legend([‘Train’, ‘Test’]) plt.grid(True, alpha=0.3, axis=’y’) plt.tight_layout() plt.show()

import numpy as np # 花びらの特徴量だけを使用（列2と3） X_2d = iris.data[:, [2, 3]] # petal length, petal width y = iris.target # データ分割 X_train_2d, X_test_2d, y_train, y_test = train_test_split( X_2d, y, test_size=0.2, random_state=42, stratify=y ) # モデル訓練 model_2d = DecisionTreeClassifier(max_depth=3, random_state=42) model_2d.fit(X_train_2d, y_train) # 評価 print(f”訓練精度: {model_2d.score(X_train_2d, y_train):.4f}”) print(f”テスト精度: {model_2d.score(X_test_2d, y_test):.4f}”) # 決定境界の可視化 def plot_decision_boundary(model, X, y): # メッシュグリッドの作成 x_min, x_max = X[:, 0].min() – 0.5, X[:, 0].max() + 0.5 y_min, y_max = X[:, 1].min() – 0.5, X[:, 1].max() + 0.5 xx, yy = np.meshgrid( np.linspace(x_min, x_max, 200), np.linspace(y_min, y_max, 200) ) # 予測 Z = model.predict(np.c_[xx.ravel(), yy.ravel()]) Z = Z.reshape(xx.shape) # プロット plt.figure(figsize=(10, 8)) plt.contourf(xx, yy, Z, alpha=0.3, cmap=’viridis’) # データポイント for i, species in enumerate(iris.target_names): mask = y == i plt.scatter( X[mask, 0], X[mask, 1], label=species, s=50, edgecolors=’black’, linewidth=1 ) plt.xlabel(‘Petal Length (cm)’) plt.ylabel(‘Petal Width (cm)’) plt.title(‘Decision Boundary – 2D Features’) plt.legend() plt.grid(True, alpha=0.3) plt.show() # 可視化 plot_decision_boundary(model_2d, X_2d, y)

from sklearn.model_selection import cross_val_score # モデルの作成 model = DecisionTreeClassifier(max_depth=3, random_state=42) # 5分割交差検証 cv_scores = cross_val_score( model, iris.data, iris.target, cv=5, # 5分割 scoring=’accuracy’ ) print(“=== 交差検証の結果 ===”) print(f”各分割の精度: {cv_scores}”) print(f”平均精度: {cv_scores.mean():.4f}”) print(f”標準偏差: {cv_scores.std():.4f}”) print(f”95%信頼区間: {cv_scores.mean():.4f} ± {1.96 * cv_scores.std():.4f}”) # 可視化 plt.figure(figsize=(10, 6)) plt.bar(range(1, 6), cv_scores, color=’steelblue’, alpha=0.7) plt.axhline(y=cv_scores.mean(), color=’red’, linestyle=’–‘, label=f’Mean: {cv_scores.mean():.4f}’) plt.xlabel(‘Fold’) plt.ylabel(‘Accuracy’) plt.title(‘Cross-Validation Scores’) plt.ylim(0.8, 1.0) plt.legend() plt.grid(True, alpha=0.3, axis=’y’) plt.show()

# データ分割（共通） X_train, X_test, y_train, y_test = train_test_split( iris.data, iris.target, test_size=0.2, random_state=42, stratify=iris.target ) # max_depthを変えて評価 max_depths = range(1, 11) train_scores = [] test_scores = [] for depth in max_depths: # モデル作成 model = DecisionTreeClassifier(max_depth=depth, random_state=42) # 訓練 model.fit(X_train, y_train) # 評価 train_scores.append(model.score(X_train, y_train)) test_scores.append(model.score(X_test, y_test)) # 可視化 plt.figure(figsize=(10, 6)) plt.plot(max_depths, train_scores, ‘o-‘, label=’Training Score’, linewidth=2) plt.plot(max_depths, test_scores, ‘s-‘, label=’Test Score’, linewidth=2) plt.xlabel(‘Max Depth’) plt.ylabel(‘Accuracy’) plt.title(‘Training vs Test Accuracy by Max Depth’) plt.legend() plt.grid(True, alpha=0.3) plt.xticks(max_depths) plt.ylim(0.8, 1.05) plt.show() # 最適なmax_depthを見つける best_depth = max_depths[np.argmax(test_scores)] best_test_score = max(test_scores) print(f”\n最適なmax_depth: {best_depth}”) print(f”そのときのテスト精度: {best_test_score:.4f}”)

# === 1. データ準備 === import numpy as np import pandas as pd import matplotlib.pyplot as plt import seaborn as sns from sklearn.datasets import load_iris from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV from sklearn.tree import DecisionTreeClassifier from sklearn.linear_model import LogisticRegression from sklearn.neighbors import KNeighborsClassifier from sklearn.svm import SVC from sklearn.metrics import accuracy_score, confusion_matrix, classification_report # データ読み込み iris = load_iris() df = pd.DataFrame(iris.data, columns=iris.feature_names) df[‘species’] = iris.target print(“=== データの基本情報 ===”) print(df.info()) print(“\n=== 基本統計量 ===”) print(df.describe()) # === 2. データ可視化 === # ペアプロット sns.pairplot(df, hue=’species’, palette=’Set2′) plt.suptitle(‘Iris Dataset – Pairplot’, y=1.02) plt.show() # 相関行列 plt.figure(figsize=(8, 6)) sns.heatmap(df.iloc[:, :-1].corr(), annot=True, cmap=’coolwarm’, center=0) plt.title(‘Feature Correlation Matrix’) plt.show() # === 3. データ分割 === X = iris.data y = iris.target X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.2, random_state=42, stratify=y ) # === 4. 複数モデルの比較 === models = { ‘DecisionTree’: DecisionTreeClassifier(random_state=42), ‘LogisticRegression’: LogisticRegression(max_iter=200, random_state=42), ‘KNN’: KNeighborsClassifier(), ‘SVM’: SVC(random_state=42) } print(“\n=== モデル比較（交差検証） ===”) cv_results = {} for name, model in models.items(): scores = cross_val_score(model, X, y, cv=5, scoring=’accuracy’) cv_results[name] = scores print(f”{name:20s}: {scores.mean():.4f} (±{scores.std():.4f})”) # === 5. 最良モデルの選択とチューニング === print(“\n=== ハイパーパラメータチューニング（DecisionTree） ===”) param_grid = { ‘max_depth’: [2, 3, 4, 5, 6, 7, 8], ‘min_samples_split’: [2, 5, 10], ‘min_samples_leaf’: [1, 2, 4] } grid_search = GridSearchCV( DecisionTreeClassifier(random_state=42), param_grid, cv=5, scoring=’accuracy’, n_jobs=-1 ) grid_search.fit(X_train, y_train) print(f”最良パラメータ: {grid_search.best_params_}”) print(f”最良スコア: {grid_search.best_score_:.4f}”) # === 6. 最終評価 === best_model = grid_search.best_estimator_ y_pred = best_model.predict(X_test) print(“\n=== 最終評価 ===”) print(f”テスト精度: {accuracy_score(y_test, y_pred):.4f}”) print(“\n分類レポート:”) print(classification_report(y_test, y_pred, target_names=iris.target_names)) # === 7. 結果の可視化 === # 混同行列 plt.figure(figsize=(8, 6)) cm = confusion_matrix(y_test, y_pred) sns.heatmap(cm, annot=True, fmt=’d’, cmap=’Blues’, xticklabels=iris.target_names, yticklabels=iris.target_names) plt.title(‘Confusion Matrix’) plt.ylabel(‘True Label’) plt.xlabel(‘Predicted Label’) plt.show() print(“\n✅ プロジェクト完了！”)

# モデルの訓練（全データで） final_model = DecisionTreeClassifier(max_depth=3, random_state=42) final_model.fit(iris.data, iris.target) def predict_iris_species(sepal_length, sepal_width, petal_length, petal_width): “”” アヤメの品種を予測する関数 Parameters: ———– sepal_length : float – がく片の長さ (cm) sepal_width : float – がく片の幅 (cm) petal_length : float – 花びらの長さ (cm) petal_width : float – 花びらの幅 (cm) Returns: ——– dict : 予測結果（品種名と確率） “”” # 入力データを配列に変換 X_new = np.array([[sepal_length, sepal_width, petal_length, petal_width]]) # 予測 prediction = final_model.predict(X_new)[0] species_name = iris.target_names[prediction] # 確率（決定木は predict_proba をサポート） probabilities = final_model.predict_proba(X_new)[0] # 結果を辞書で返す result = { ‘species’: species_name, ‘class_id’: int(prediction), ‘probabilities’: { iris.target_names[i]: float(prob) for i, prob in enumerate(probabilities) } } return result # === テスト === print(“=== テスト 1: Setosaっぽい花 ===”) result1 = predict_iris_species(5.1, 3.5, 1.4, 0.2) print(f”予測品種: {result1[‘species’]}”) print(“確率:”) for species, prob in result1[‘probabilities’].items(): print(f” {species:12s}: {prob:.4f} ({prob*100:.2f}%)”) print(“\n=== テスト 2: Virginicaっぽい花 ===”) result2 = predict_iris_species(6.5, 3.0, 5.5, 2.0) print(f”予測品種: {result2[‘species’]}”) print(“確率:”) for species, prob in result2[‘probabilities’].items(): print(f” {species:12s}: {prob:.4f} ({prob*100:.2f}%)”)