主函数
import numpy as np
def normalization(data: np.array,y_dimenion:int):
col, row = data.shape
m = []
s = []
for i in range(row-y_dimenion):
m.append(np.sum(data[:, i])/col)
for i in range(row-y_dimenion):
s.append(np.sum(np.abs(data[:, i]))/col)
for i in range(row-y_dimenion):
data[:, i] -= m[i]
data[:, i] /= s[i]
return data
- 说明:
- data以numpy矩阵的形式输入
- 一行为一个数据元组
- y_dimension为分类器的类别,放在数据元组的末尾
- y_dimension可以理解为最后i列数据不标准化