如何编写实用的机器学习代码及实战案例

X[col].fillna(X[col].median, inplace=True)
for col in categorical_features:
if X[col].isnull.sum > 0:
X[col].fillna(X[col].mode[0] if not X[col].mode.empty else ‘Missing’, inplace=True)
# 创建预处理管道
numerical_transformer = Pipeline(steps=[
(‘scaler’, StandardScaler)
])
categorical_transformer = Pipeline(steps=[
(‘onehot’, OneHotEncoder(handle_unknown=’ignore’, sparse_output=False))
])
self.preprocessor = ColumnTransformer(
transformers=[
(‘num’, numerical_transformer, numerical_features),
(‘cat’, categorical_transformer, categorical_features)
])
self.feature_names = numerical_features + categorical_features
return X, y
def build_model(self):
构建并训练模型
# 创建完整的管道
model_pipeline = Pipeline(steps=[
(‘preprocessor’, self.preprocessor),
(‘regressor’, GradientBoostingRegressor(
n_estimators=100,
learning_rate=0.1,
max_depth=3,
random_state=42
))
])
return model_pipeline
def train(self, X, y, test_size=0.2):
训练模型并评估性能
# 分割数据集
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=test_size, random_state=42
# 构建和训练模型
self.model = self.build_model
self.model.fit(X_train, y_train)
# 预测和评估
y_pred = self.model.predict(X_test)
mse = mean_squared_error(y_test, y_pred)
rmse = np.sqrt(mse)
r2 = r2_score(y_test, y_pred)
print(f”模型评估结果:”)
print(f”RMSE: {rmse:.2f}”)
print(f”R² Score: {r2:.4f}”)
# 交叉验证
cv_scores = cross_val_score(self.model, X_train, y_train,
cv=5, scoring=’neg_mean_squared_error’)
cv_rmse = np.sqrt(-cv_scores.mean)
print(f”5折交叉验证 RMSE: {cv_rmse:.2f}”)
return X_test, y_test, y_pred
def feature_importance(self, X):
分析特征重要性
if hasattr(self.model.named_steps[‘regressor’], ‘feature_importances_’):
# 获取特征名称（考虑one-hot编码后的特征）
preprocessor = self.model.named_steps[‘preprocessor’]
feature_names = []
# 数值型特征名称
feature_names.extend(X.select_dtypes(include=[‘int64’, ‘float64’]).columns.tolist)
# 类别型特征名称（one-hot编码后）
categorical_features = X.select_dtypes(include=[‘object’]).columns.tolist
for feature in categorical_features:
categories = preprocessor.named_transformers_[‘cat’].named_steps[‘onehot’].categories_[
categorical_features.index(feature)
for category in categories:
feature_names.append(f”{feature}_{category}”)
importances = self.model.named_steps[‘regressor’].feature_importances_
feature_imp_df = pd.DataFrame({
‘feature’: feature_names[:len(importances)],
‘importance’: importances
}).sort_values(‘importance’, ascending=False)
# 可视化前20个重要特征
plt.figure(figsize=(10, 8))
sns.barplot(data=feature_imp_df.head(20), x=’importance’, y=’feature’)
plt.title(‘Top 20 Feature Importance’)
plt.tight_layout
plt.show
return feature_imp_df
def save_model(self, file_path):
保存训练好的模型
if self.model is not None:
joblib.dump(self.model, file_path)
print(f”模型已保存至: {file_path}”)
def load_model(self, file_path):
加载已保存的模型
self.model = joblib.load(file_path)
print(f”模型已从 {file_path} 加载”)
def main:
主函数
演示完整工作流程
predictor = HousePricePredictor
# 加载数据（这里使用示例数据路径，实际使用时需要修改）
try:
data = predictor.load_data(‘house_prices.csv’)
except FileNotFoundError:
print(“请提供正确的数据文件路径”)
return
# 数据预处理
X, y = predictor.preprocess_features(data)
# 训练模型
X_test, y_test, y_pred = predictor.train(X, y)
# 分析特征重要性
feature_importance_df = predictor.feature_importance(X)
print(”
前10个重要特征:”)
print(feature_importance_df.head(10))
# 保存模型
predictor.save_model(‘house_price_model.pkl’)
if __name__ == “__main__”:
main


project_name=python_房价预测
filename=requirements.txt
title=项目依赖文件
entrypoint=false
runnable=false
project_final_file=true
pandas>=1.5.0
numpy>=1.21.0
scikit-learn>=1.0.0
matplotlib>=3.5.0
seaborn>=0.11.0
joblib>=1.2.0