1. 安装依赖

首先，确保安装了所需的库：

pip install yfinance pandas numpy statsmodels matplotlib

2. 策略框架结构

我们将创建一个策略框架，允许用户定义不同的策略，并使用 yfinance 获取数据。以下是一个简单的实现：

import yfinance as yf
import pandas as pd
import numpy as np
import statsmodels.api as sm
import matplotlib.pyplot as plt

class Strategy:
    def __init__(self, tickers, start_date, end_date):
        self.tickers = tickers
        self.start_date = start_date
        self.end_date = end_date
        self.data = self.get_data()

    def get_data(self):
        """Fetch historical data from Yahoo Finance."""
        data = yf.download(self.tickers, start=self.start_date, end=self.end_date)
        return data['Close']

    def calculate_returns(self):
        """Calculate daily returns."""
        returns = self.data.pct_change().dropna()
        return returns

    def backtest(self):
        """Run the backtest for the defined strategy."""
        raise NotImplementedError("Backtest method not implemented.")

class MeanReversionStrategy(Strategy):
    def __init__(self, tickers, start_date, end_date, lookback=20, entry_zscore=1):
        super().__init__(tickers, start_date, end_date)
        self.lookback = lookback
        self.entry_zscore = entry_zscore

    def backtest(self):
        returns = self.calculate_returns()
        signals = (returns.rolling(self.lookback).mean() - returns) / returns.rolling(self.lookback).std()
        
        long_positions = signals < -self.entry_zscore
        short_positions = signals > self.entry_zscore
        
        positions = pd.DataFrame(index=returns.index, columns=returns.columns).fillna(0)
        positions[long_positions] = 1
        positions[short_positions] = -1
        
        daily_returns = (positions.shift() * returns).sum(axis=1)
        cumulative_returns = (1 + daily_returns).cumprod() - 1
        
        return cumulative_returns

class MomentumStrategy(Strategy):
    def __init__(self, tickers, start_date, end_date, lookback=20):
        super().__init__(tickers, start_date, end_date)
        self.lookback = lookback

    def backtest(self):
        returns = self.calculate_returns()
        signals = returns.rolling(self.lookback).mean()
        
        positions = pd.DataFrame(index=returns.index, columns=returns.columns).fillna(0)
        positions[signals > 0] = 1  # Long positions
        
        daily_returns = (positions.shift() * returns).sum(axis=1)
        cumulative_returns = (1 + daily_returns).cumprod() - 1
        
        return cumulative_returns

# Example usage
if __name__ == "__main__":
    tickers = ['AAPL', 'MSFT', 'GOOGL']  # Example tickers
    start_date = '2020-01-01'
    end_date = '2023-01-01'

    # Mean Reversion Strategy
    mean_reversion = MeanReversionStrategy(tickers, start_date, end_date)
    mean_reversion_returns = mean_reversion.backtest()
    
    # Momentum Strategy
    momentum = MomentumStrategy(tickers, start_date, end_date)
    momentum_returns = momentum.backtest()

    # Plotting results
    plt.figure(figsize=(12, 6))
    plt.plot(mean_reversion_returns, label='Mean Reversion Strategy')
    plt.plot(momentum_returns, label='Momentum Strategy')
    plt.title('Strategy Backtest Results')
    plt.xlabel('Date')
    plt.ylabel('Cumulative Returns')
    plt.legend()
    plt.show()

3. 说明

• Strategy 类: 这是一个基类，负责获取数据和计算收益。
• MeanReversionStrategy 类: 继承自 Strategy，实现了均值回归策略的回测。
• MomentumStrategy 类: 继承自 Strategy，实现了动量策略的回测。
• 数据获取: 使用 yfinance 获取指定资产的历史数据。
• 回测: 每个策略都有自己的 backtest 方法，计算策略的累积收益。

4. 扩展

您可以根据需要扩展更多策略，只需继承 Strategy 类并实现 backtest 方法。此框架为不同的资产和策略提供了灵活的回测能力。