用代码思考 用数据决策

# 双均线策略（简化版）
def dual_ma_strategy(df, short=5, long=20):
    df['ma_short'] = df['close'].rolling(short).mean()
    df['ma_long'] = df['close'].rolling(long).mean()
    df['signal'] = 0
    df.loc[df['ma_short'] > df['ma_long'], 'signal'] = 1
    df['position'] = df['signal'].diff()
    return df

# 布林带均值回归策略
def bollinger_reversion(df, window=20, std=2):
    df['mid'] = df['close'].rolling(window).mean()
    df['std'] = df['close'].rolling(window).std()
    df['upper'] = df['mid'] + std * df['std']
    df['lower'] = df['mid'] - std * df['std']
    # 价格触及下轨买入，上轨卖出
    df['signal'] = 0
    df.loc[df['close'] < df['lower'], 'signal'] = 1
    df.loc[df['close'] > df['upper'], 'signal'] = -1
    return df

# 动量因子选股策略
def momentum_factor(df, lookback=60):
    # 计算过去N日的累计收益率
    df['momentum'] = df['close'].pct_change(lookback)
    # 选取动量排名前20%的股票
    df['rank'] = df['momentum'].rank(pct=True)
    df['signal'] = (df['rank'] > 0.8).astype(int)
    return df