代码

import numpy as np
def normalization(x):
    """"
    归一化到区间{0,1]
    返回副本
    """
    _range = np.max(x) - np.min(x)
    return (x - np.min(x)) / _range
def standardization(x):
    """"
     将输入x 正态标准化  (x - mu) / sigma   ~   N(0,1)
     返回副本
    """
    mu = np.mean(x, axis=0)
    sigma = np.std(x, axis=0)
    return (x - mu) / sigma
YUAN = np.random.randint(3, 70, size=5)
a = normalization(YUAN)
b = standardization(YUAN)
print(YUAN, "normalization:", a)
print(YUAN is a)
print(YUAN, "standardization:", b)
print(YUAN is b)

结果

[33 50 60 12 32] normalization: [0.4375     0.79166667 1.         0.         0.41666667]

False

[33 50 60 12 32] standardization: [-0.26647587  0.76308999  1.36871697 -1.53829253 -0.32703857]

False

sklearn.preprocessing.scale 能沿某个轴标准化

import numpy as np
from sklearn import  preprocessing
def standardization(x):
    """"
     将输入x 正态标准化  (x - mu) / sigma   ~   N(0,1)
     返回副本
    """
    mu = np.mean(x, axis=0)
    sigma = np.std(x, axis=0)
    return (x - mu) / sigma
YUAN = np.random.randint(3, 70, size=5)
a = standardization(YUAN)
b = preprocessing.scale(YUAN)
print(YUAN, "normalization:", a)
print(YUAN, "preprocessing.scale:", b)
print(a == b)

结果

[ 8 14 64 51 61] normalization: [-1.32656065 -1.07468204  1.02430632  0.47856935  0.89836702]
[ 8 14 64 51 61] preprocessing.scale: [-1.32656065 -1.07468204  1.02430632  0.47856935  0.89836702]
[ True  True  True  True  True]