将label矩阵转变为0-1矩阵

机器学习中，经常需要将label matrix （0-1）与 label 向量（为类别标签）进行相互转换。前者是在机器学习的模型中广泛使用，后者在分类（聚类）结果的评价中使用。下面提供python和matlab的两个平台上面的函数代码。

1、python code

import numpy as np
# label vector converted to label matrix
def label2matrix(label):
    label = np.array(label)
    uq_la = np.unique(label)
    c = uq_la.shape[0]
    n = label.shape[0]
    label_mat = np.zeros((n,c))
    for i in range(c):
        index = (label == i+1)
        label_mat[index,i]=1
    return label_mat

# label matrix converted to label vector
def matrix2label(label_mat):
    label_mat = np.array(label_mat)
    n,c = label_mat.shape
    label = np.zeros((n,))
    for i in range(c):
        label[label_mat[:,i]==1] = i+1
    return label

2、matlab

function label_mat = label2matrix(label)
uq_label = unique(label);
n = length(label);
c = length(uq_label);
label_mat = zeros(n,c);
for i = 1:c
    index = label == i;
    label_mat(index,i) = 1;
end
end

function label_vec = matrix2label(matrix)
   [n,c] = size(matrix);
   label_vec = zeros(1,n)
   for i = 1:c
       label_vec(matrix(:,i)==1) = i
   end
end