Python数据标准化

1.最小最大标准化

def MinMaxNorm(x):
    # 标准化
    min_val = np.nanmin(x, axis=0)
    max_val = np.nanmax(x, axis=0)
    for i in range(len(x)):
        x[i] = (x[i] - min_val) / (max_val - min_val)
    return x

2.Z-Score标准化

# 标准化
def robust_zscore(x: pd.Series, zscore=False):
    # https://en.wikipedia.org/wiki/Median_absolute_deviation
    # 删除以异常值
    x = x - x.median()
    mad = x.abs().median()
    x = np.clip(x / mad / 1.4826, -3, 3)
    
    # 标准化
    if zscore:
        x -= x.mean()
        x /= x.std()
    return x

3.Z-Score另一种表达方式

# Z值标准化
def standardize_z(dt):
    mean = dt.mean()     #  截面数据均值
    std = dt.std()       #  截面数据标准差
    return (dt - mean)/std

# norm_df = pd.DataFrame()
# norm_df["Date"] = process_data["Date"]
# norm_df["Code"] = process_data["Code"]

norm_df = pd.DataFrame()
for day in process_data.Date.unique():
    split_df = process_data[process_data["Date"]==day]
    split_df.iloc[:,2:] = (split_df.iloc[:,2:] - split_df.iloc[:,2:].mean())/split_df.iloc[:,2:].std()
    norm_df = norm_df.append(split_df)