머신 러닝 (14) Feature Engineering regression

 import pandas as pd

import numpy as np

from sklearn.datasets import load_diabetes

from sklearn.model_selection import train_test_split

from sklearn.preprocessing import StandardScaler

from sklearn.feature_selection import SelectFromModel

from sklearn.linear_model import LassoCV, Ridge

from sklearn.pipeline import Pipeline

from sklearn.metrics import r2_score

# 데이터 로드

diabetes = load_diabetes()

X = pd.DataFrame(diabetes.data, columns=diabetes.feature_names)

y = diabetes.target

print(f"데이터 크기: {X.shape}")

print(f"피처 목록: {list(diabetes.feature_names)}")

diabetes = load_diabetes(return_X_y=False)

X = pd.DataFrame(diabetes.data, columns=diabetes.feature_names)

y = diabetes.target

print(f"데이터 크기: {X.shape}")

print(f"피처 목록: {list(diabetes.feature_names)}")

y

y.dtype

X

# 데이터 분리

X_train, X_test, y_train, y_test = train_test_split(

    X, y, test_size=0.2, random_state=42

)

# Pipeline 구축

pipe_lasso = Pipeline([

    ('scaler', StandardScaler()),

    ('feature_selection', SelectFromModel(LassoCV(cv=5, random_state=42))),

    ('regressor', Ridge(random_state=42))

])

print("Pipeline 구성 완료")

print(pipe_lasso)

# 학습

print("Training Pipeline with LassoCV Selection...")

pipe_lasso.fit(X_train, y_train)

print("학습 완료!")

# 선택된 피처 확인

selector = pipe_lasso.named_steps['feature_selection']

selected_indices = selector.get_support(indices=True)

selected_features = np.array(diabetes.feature_names)[selected_indices]

print(f"선택된 피처 개수: {len(selected_features)}개 / {len(diabetes.feature_names)}개")

print(f"선택된 피처 목록: {list(selected_features)}")

# LassoCV 모델 추출

lasso_model = selector.estimator_

print(f"최적 Alpha: {lasso_model.alpha_:.6f}")

print(f"\n각 피처별 계수:")

coef_df = pd.DataFrame({

    'Feature': diabetes.feature_names,

    'Coefficient': lasso_model.coef_,

    'Abs_Coef': np.abs(lasso_model.coef_),

    'Selected': selector.get_support()

}).sort_values('Abs_Coef', ascending=False)

print(coef_df)

# 예측 및 평가

y_pred = pipe_lasso.predict(X_test)

r2 = r2_score(y_test, y_pred)

print(f"Pipeline R2 Score: {r2:.4f}")

# 모든 피처 사용 (피처 선택 없음)

pipe_all = Pipeline([

    ('scaler', StandardScaler()),

    ('regressor', Ridge(random_state=42))

])

pipe_all.fit(X_train, y_train)

y_pred_all = pipe_all.predict(X_test)

r2_all = r2_score(y_test, y_pred_all)

# 결과 비교

print("성능 비교:")

print(f"모든 피처 사용 (10개): R2 = {r2_all:.4f}")

print(f"선택된 피처 ({len(selected_features)}개): R2 = {r2:.4f}")

print(f"성능 차이: {r2 - r2_all:+.4f}")

print(f"\n피처 감소율: {(10 - len(selected_features)) / 10 * 100:.1f}%")

import matplotlib.pyplot as plt

# 실제값 vs 예측값 시각화

plt.figure(figsize=(10, 6))

plt.scatter(y_test, y_pred, alpha=0.6, edgecolors='k')

plt.plot([y_test.min(), y_test.max()], [y_test.min(), y_test.max()],

         'r--', lw=2, label='Perfect Prediction')

plt.xlabel('Actual Values')

plt.ylabel('Predicted Values')

plt.title(f'Actual vs Predicted (R2={r2:.4f})')

plt.legend()

plt.grid(True, alpha=0.3)

plt.show()