需要遍历每列并找到最大值和最小值

Question

import numpy as np

x = np.loadtxt('xdata.txt', dtype=float)
y = np.loadtxt('ydata.txt', dtype=float)

normalX = []
normalY = []

for column in x:
    i = 0
    while i <=17:
        xmax = max(column[i])
        xmin = min(column[i])
        normalx = (?-xmin)/(xmax-xmin)
        normalX.append(normalx)
        i += 1
    else:
        break

I have a 148 by 17 matrix that I import and I want to normalize the data. I am trying to iterate through each column and find the max and min but my code so far results in an "TypeError: 'numpy.float64' object is not iterable". Also what should ? be if I want to have it be the element in the column.

Instead of a big 148x17 matrix ill put a 4x4.

1.61  125  13   933.57 
1.95  135  29   1357.77 
1.91  135  28   1728 
2.2   137  46   1828.05 

First column max would be 2.2, min = 1.61 etc.

Answer 1

In your code, you are first accessing for columns in x this causes columns to be the row in x , then later on inside the for loop you are trying to access columns[i] , this would return the element in i position in that row.

You can use np.amax(x, axis=0) to get the maximum of each column , this would inturn return a list of maximum for each column.

Similarly, you can also use np.amin(x, axis=0) .

Another issue is that if your matrix is of size 148x17 , in your while loop you should only be checking till i<17 not i<=17 . You can also you enumerate(column) along with for loop for same , the syntax would be - for i, val in enumerate(column) .

Also, I am guessing you need to normalize all the values, so you need to create a list inside the first for loop and keep inserting to that list and then at the end , add that list to normalX .

Example code -

xmax = np.amax(x, axis=0)
xmin = np.amin(x, axis=0)

for column in x:
    tnormalX = []
    for i, val in enumerate(column):
        normalx = (val-xmin[i])/(xmax[i]-xmin[i])
        tnormalX.append(normalx)
        i += 1
    normalX.append(tnormalX)