Processing of normalized eigenvalue of "machine learning"

# -*- coding:utf-8 -*-__author__ =  ' Ghostviper ' "" "Normalized Eigenvalue" "" from numpy  Import *def autonorm (DataSet):     minvals = dataset.min (0)      maxvals = dataset.max (0)     ranges = maxVals -  Minvals    normdataset = zeros (Shape (dataSet))     m =  dataset.shape[0]    normdataset = dataset - tile (minVals,  ( m, 1))     normdataset = normdataset / tile (ranges,  (m,  1))     return normDataSet, ranges, minValsif __name__ ==  "__ Main__ ":     dataset = array ([        [ 0.1, 12345, 23],        [-1.2, 456431, 46],         [0.99, 23332, 89],        [1.3,  97653, 123],        [2, 10900, 23],         [1, 54612, 9],    ])      Normdataset, ranges, minvals = autonorm (DataSet)

Output Result:

Array ([[ 0.40625   ,  0.00324332,  0.12280702],        [ 0.        ,  1.         ,  0.3245614 ],       [  0.684375  ,  0.02790378,  0.70175439],        [ 0.78125   ,  0.19471821,  1.         ],       [ 1.         ,  0.        ,  0.12280702],        [ 0.6875    ,  0.09811214,   0.        ]]) array ([  3.20000000e+00,    4.45531000e+05,   1.14000000E+02]) Array ([ -1.20000000e+00,   1.09000000e+04,   9.00000000e+00]) 

Algorithm Core: (Data Set-minimum feature dataset)/(Maximum feature-minimum feature) dataset

Purpose: Used to deal with different groups of characteristics of large differences in data conditions

This article is from the "Ghostviper" blog, make sure to keep this source http://ghostviper.blog.51cto.com/3396458/1865053

Processing of normalized eigenvalue of "machine learning"