电力负荷预测案例之项目搭建

import os
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import logging
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_absolute_error, mean_squared_error

# 1. 项目目录结构
print("=== 1. 项目目录结构 ===")

def create_project_structure():
    """创建项目目录结构"""
    base_dir = 'load_predict_project'
    
    directories = [
        'data',      # 放数据
        'log',       # 存放日志
        'model',     # 保存模型文件
        'src',       # 主业务逻辑
        'utils'      # 工具包
    ]
    
    for dir_name in directories:
        dir_path = os.path.join(base_dir, dir_name)
        os.makedirs(dir_path, exist_ok=True)
        print(f"  创建目录: {dir_path}")
    
    print(f"\n项目目录结构:")
    print(f"  {base_dir}/")
    for dir_name in directories:
        print(f"  ├── {dir_name}/")
    
    return base_dir

create_project_structure()

# 2. 日志工具类
print("\n=== 2. 日志工具类 ===")

class LogUtils:
    """日志工具类"""
    
    def __init__(self, log_file='log/project.log'):
        """
        初始化日志
        Args:
            log_file: 日志文件路径
        """
        self.logger = logging.getLogger('power_load_prediction')
        self.logger.setLevel(logging.DEBUG)
        
        # 避免重复添加handler
        if not self.logger.handlers:
            # 创建文件handler
            os.makedirs(os.path.dirname(log_file), exist_ok=True)
            file_handler = logging.FileHandler(log_file, encoding='utf-8')
            file_handler.setLevel(logging.DEBUG)
            
            # 创建控制台handler
            console_handler = logging.StreamHandler()
            console_handler.setLevel(logging.INFO)
            
            # 创建格式器
            formatter = logging.Formatter(
                '%(asctime)s - %(name)s - %(levelname)s - %(message)s'
            )
            file_handler.setFormatter(formatter)
            console_handler.setFormatter(formatter)
            
            # 添加handler
            self.logger.addHandler(file_handler)
            self.logger.addHandler(console_handler)
    
    def info(self, message):
        """记录info日志"""
        self.logger.info(message)
    
    def error(self, message):
        """记录error日志"""
        self.logger.error(message)
    
    def warning(self, message):
        """记录warning日志"""
        self.logger.warning(message)

# 测试日志工具
log_utils = LogUtils()
log_utils.info("项目启动")
log_utils.info("开始加载数据")

# 3. 模拟数据
print("\n=== 3. 模拟数据 ===")

def generate_power_load_data(n_days=30):
    """生成电力负荷数据"""
    np.random.seed(42)
    
    dates = pd.date_range('2024-01-01', periods=n_days*24, freq='h')
    power_load = []
    
    for date in dates:
        hour = date.hour
        dayofweek = date.dayofweek
        month = date.month
        
        # 基础负荷
        if 0 <= hour < 6:
            base_load = 500
        elif 6 <= hour < 9:
            base_load = 800
        elif 9 <= hour < 17:
            base_load = 1000
        elif 17 <= hour < 21:
            base_load = 1200
        else:
            base_load = 700
        
        # 周末调整
        if dayofweek >= 5:
            base_load *= 0.8
        
        # 季节调整
        if month in [12, 1, 2]:
            base_load *= 1.2
        elif month in [6, 7, 8]:
            base_load *= 1.1
        
        load = base_load + np.random.normal(0, 50)
        power_load.append(max(100, load))
    
    data = pd.DataFrame({
        'time': dates,
        'power_load': power_load
    })
    
    return data

data = generate_power_load_data(n_days=30)
log_utils.info(f"数据加载完成，形状: {data.shape}")
print(f"数据形状: {data.shape}")

# 4. 特征工程
print("\n=== 4. 特征工程 ===")

def extract_features(data):
    """提取特征"""
    data = data.copy()
    
    # 基础时间特征
    data['hour'] = data['time'].dt.hour
    data['day'] = data['time'].dt.day
    data['month'] = data['time'].dt.month
    data['dayofweek'] = data['time'].dt.dayofweek
    
    # 衍生特征
    data['is_weekend'] = data['dayofweek'].isin([5, 6]).astype(int)
    data['is_daytime'] = ((data['hour'] >= 6) & (data['hour'] < 18)).astype(int)
    data['quarter'] = data['time'].dt.quarter
    
    # 上个小时负荷
    data['prev_load'] = data['power_load'].shift(1)
    data['prev_load'].fillna(method='bfill', inplace=True)
    
    return data

data_with_features = extract_features(data)
log_utils.info(f"特征工程完成，特征数: {data_with_features.shape[1]}")
print(f"特征数: {data_with_features.shape[1]}")

# 5. 准备训练数据
print("\n=== 5. 准备训练数据 ===")

def prepare_training_data(data):
    """准备训练数据"""
    features = ['hour', 'day', 'month', 'dayofweek', 'is_weekend', 
                'is_daytime', 'quarter', 'prev_load']
    
    X = data[features].values
    y = data['power_load'].values
    
    return X, y, features

X, y, features = prepare_training_data(data_with_features)
log_utils.info(f"训练数据准备完成: X={X.shape}, y={y.shape}")
print(f"特征: {features}")
print(f"X形状: {X.shape}")
print(f"y形状: {y.shape}")

# 6. 划分训练测试集
print("\n=== 6. 划分训练测试集 ===")
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42, shuffle=False
)

log_utils.info(f"训练集: {X_train.shape}")
log_utils.info(f"测试集: {X_test.shape}")
print(f"训练集: {X_train.shape}")
print(f"测试集: {X_test.shape}")

# 7. 训练模型
print("\n=== 7. 训练模型 ===")

def train_model(X_train, y_train):
    """训练模型"""
    try:
        from sklearn.ensemble import GradientBoostingRegressor
        
        model = GradientBoostingRegressor(
            n_estimators=100,
            max_depth=6,
            learning_rate=0.1,
            random_state=42
        )
        
        model.fit(X_train, y_train)
        log_utils.info("模型训练完成")
        print("模型训练完成")
        
        return model
    except Exception as e:
        log_utils.error(f"模型训练失败: {e}")
        raise

model = train_model(X_train, y_train)

# 8. 评估模型
print("\n=== 8. 评估模型 ===")

def evaluate_model(model, X_test, y_test):
    """评估模型"""
    try:
        y_pred = model.predict(X_test)
        
        mae = mean_absolute_error(y_test, y_pred)
        mse = mean_squared_error(y_test, y_pred)
        rmse = np.sqrt(mse)
        
        log_utils.info(f"MAE: {mae:.4f}")
        log_utils.info(f"MSE: {mse:.4f}")
        log_utils.info(f"RMSE: {rmse:.4f}")
        
        print(f"MAE: {mae:.4f}")
        print(f"MSE: {mse:.4f}")
        print(f"RMSE: {rmse:.4f}")
        
        return y_pred
    except Exception as e:
        log_utils.error(f"模型评估失败: {e}")
        raise

y_pred = evaluate_model(model, X_test, y_test)

# 9. 保存模型
print("\n=== 9. 保存模型 ===")

def save_model(model, model_path='model/power_load_model.pkl'):
    """保存模型"""
    try:
        os.makedirs(os.path.dirname(model_path), exist_ok=True)
        
        import joblib
        joblib.dump(model, model_path)
        
        log_utils.info(f"模型已保存: {model_path}")
        print(f"模型已保存: {model_path}")
    except Exception as e:
        log_utils.error(f"模型保存失败: {e}")
        raise

save_model(model)

# 10. 可视化预测结果
print("\n=== 10. 可视化预测结果 ===")
plt.figure(figsize=(14, 6))
plt.plot(y_test, label='真实值', marker='o', alpha=0.7)
plt.plot(y_pred, label='预测值', alpha=0.7)
plt.xlabel('时间步')
plt.ylabel('电力负荷')
plt.title('电力负荷预测结果')
plt.legend()
plt.grid(True, alpha=0.3)
plt.savefig('log/prediction_result.png', dpi=100)
log_utils.info("预测结果图已保存")
print("预测结果图已保存")

# 11. 完整项目流程
print("\n=== 11. 完整项目流程 ===")

def project_pipeline():
    """完整项目流程"""
    print("=" * 50)
    print("电力负荷预测 - 项目流程")
    print("=" * 50)
    
    print("\n1. 项目搭建")
    print("   创建目录结构: data, log, model, src, utils")
    
    print("\n2. 日志配置")
    print("   配置日志工具类")
    print("   记录项目进展和错误")
    
    print("\n3. 数据加载")
    print("   加载电力负荷数据")
    
    print("\n4. 特征工程")
    print("   从时间提取特征")
    print("   生成多变量特征")
    
    print("\n5. 模型训练")
    print("   划分训练测试集")
    print("   训练XGBoost模型")
    
    print("\n6. 模型评估")
    print("   MAE、MSE、RMSE")
    
    print("\n7. 模型保存")
    print("   保存到model目录")
    
    print("\n8. 预测可视化")
    print("   绘制预测结果图")
    
    print("\n" + "=" * 50)
    print("项目流程完成!")
    print("=" * 50)

project_pipeline()

# 12. 项目说明
print("\n=== 12. 项目说明 ===")
print("""
项目说明:

目标: 基于历史数据训练XGBoost模型
      实现多变量单步电力负荷预测

特点:
- 多变量: 多列特征（小时、月、上个小时负荷等）
- 单步: 标签列只有一列（电力负荷）

目录结构:
├── data/       # 放数据
├── log/        # 存放日志
├── model/      # 保存模型文件
├── src/        # 主业务逻辑
└── utils/      # 工具包

开发规范:
- 命名规范，不要有中文
- 工具包可复用
- 使用try-except记录错误
""")