当前位置：网站首页>How to standardize multidimensional matrix (based on numpy)

How to standardize multidimensional matrix (based on numpy)

2022-04-23 07:17:00 【Breeze_】

step1. Load or generate a matrix 

import numpy as np
#  Load from a file 
matrix = np.load("xx.npy")
#  Randomly generate one 2x3x4 Dimensional 1 To 100 The random matrix of 
X = np.random.randint(1,100,(2,3,4))

step2. Calculation mean and std

means = np.mean(X,axis=(0,2)) # Calculate the mean , First calculate the number 0 Dimensional , Post calculation section 2 Dimensional , Get one shape by 3, Matrix 
stds = np.std(X,axis = (0,2)) # Calculate variance , Get one shape by 3, Matrix 
means = means.reshape(1,-1,1) # Convert to and X A matrix of the same dimension , here reshape The matrix dimension after is (1,3,1)
stds = stds.reshape(1,-1,1)  # ditto 
X = (X-means)/stds

版权声明
本文为[Breeze_]所创，转载请带上原文链接，感谢
https://yzsam.com/2022/04/202204230610323183.html

边栏推荐

猜你喜欢

随机推荐