python提取列表中非零元素的索引--多标签分类输出相应类别

在多标签分类学习中,神经网络的输出通常是以0,1,想要将预测结果翻译成真实标签就需要获取网络输出中"1"元素的索引值,然后相应的提取出标签.

 

import numpy as np
labellist = ['a','b','c','d','e','f','g']
# 输出为:
outlist = [1,0,0,0,1,0,0]
# 对应的输出标签应该是:['a','e']
nozeroindex = []
outlabelist = []
for x in range(len(outlist)):
    nozeroindex = np.nonzero(outlist)[0]
    for x in nozeroindex:
        label = labellist[x]
        outlabelist.append(label)
    print(outlabelist)