How to turn a multi-label np.array to one vs all?-CodePudding

I have an np.array with several labels, let's say label1, ..., label10.

Given a specific label, let's say label1. I want to turn it to a binary vector where y'[j] == 1 iff y[j] == label1

How to do that?

CodePudding user response：

Alright, I found a fine way:

import numpy as np
label = 'label1'
tobinary = lambda val: 1 if val == label else 0

vfunc = np.vectorize(tobinary)
vfunc(y)

relying on this SO answer

CodePudding user response：

Simple comparison:

>>> np.array([f'label{i}' for i in range(10)])
array(['label0', 'label1', 'label2', 'label3', 'label4', 'label5',
       'label6', 'label7', 'label8', 'label9'], dtype='<U6')
>>> ar = np.array([f'label{i}' for i in range(10)])
>>> ar == 'label1'
array([False,  True, False, False, False, False, False, False, False,
       False])

If you want a binary vector:

>>> (ar == 'label1').astype(int)
array([0, 1, 0, 0, 0, 0, 0, 0, 0, 0])