题解 | 特征扩展实现

import numpy as np


def feature_scaling(data):
    # 补全代码
    data = np.array(data, dtype=float)  # 确保是浮点数运算

    # 标准化缩放
    mean = data.mean(axis=0)
    std = data.std(axis=0)
    z = np.zeros_like(data, dtype=float)
    mask_std = std != 0
    z[:, mask_std] = (data[:, mask_std] - mean[mask_std]) / std[mask_std]

    # 最小-最大标准化缩放
    minv = data.min(axis=0)
    maxv = data.max(axis=0)
    rng = maxv - minv
    mm = np.zeros_like(data, dtype=float)
    mask_rng = rng != 0
    mm[:, mask_rng] = (data[:, mask_rng] - minv[mask_rng]) / rng[mask_rng]

    # 转成列表型数据并保留四位有效数字
    z_list = np.round(z, 4).tolist()
    mm_list = np.round(mm, 4).tolist()

    return z_list, mm_list


# 主程序
if __name__ == "__main__":
    # 输入数组
    data = input()

    # # 处理输入
    # import ast
    # data = ast.literal_eval(data)
    data = eval(data)

    # 调用函数计算
    output = feature_scaling(data)

    # 输出结果
    print(output)