IC的计算公式：

# 计算Rank IC的Python示例
import pandas as pd
from scipy.stats import spearmanr

def calc_rank_ic(factor_values, forward_returns):
    """
    计算Rank IC
    factor_values: 因子值序列
    forward_returns: 未来N日收益序列
    """
    ic, p_value = spearmanr(factor_values, forward_returns)
    return ic, p_value

# 月度IC计算示例
monthly_ic = []
for month in range(12):
    ic, _ = calc_rank_ic(
        factor_data[month], 
        return_data[month]
    )
    monthly_ic.append(ic)

print(f"平均IC: {np.mean(monthly_ic):.4f}")
print(f"IC标准差: {np.std(monthly_ic):.4f}")

分组回测代码框架：

def factor_group_backtest(factor_df, return_df, groups=10):
    """
    因子分组回测
    factor_df: 因子值矩阵，行为时间，列为股票
    return_df: 未来收益矩阵
    """
    results = {}
    
    for date in factor_df.index:
        # 获取当期因子值
        factors = factor_df.loc[date].dropna()
        
        # 分组
        labels = pd.qcut(factors, groups, labels=False)
        
        # 计算每组收益
        group_returns = return_df.loc[date][factors.index]
        for g in range(groups):
            mask = labels == g
            if g not in results:
                results[g] = []
            results[g].append(group_returns[mask].mean())
    
    # 计算累计收益
    cum_returns = {}
    for g in results:
        cum_returns[g] = (1 + pd.Series(results[g])).cumprod()
    
    return cum_returns