如何从标签向量生成标签矩阵? [英] How to generate label matrix from a label vector?

问题描述

我的标签向量如下:

   y=[3 1 5 3 4 2];

是否有任何有效的方法来生成以下标签矩阵?

Is there any efficient way to generate the following label matrix?

    [0 0 1 0 0;
     1 0 0 0 0;
     0 0 0 0 1;
     0 0 1 0 0;
     0 0 0 1 0;
     0 1 0 0 0;]

更新: 这篇文章和这篇文章都是很好的答案.使用@Nras提供的脚本，以下操作可以处理丢失的标签:

UPDATED: This post and this post are both good answers. Using the scripts provided by @Nras, the following is to handle the missing labels:

    Y=[3 1 5 3 4 2];
    labels=unique(Y);
    [~,indexes]=ismember(Y,labels);
    rows = 1:length(Y); %// row indx
    T = zeros(length(Y),length(unique(indexes))); %// A matrix full of zeros
    T(sub2ind(size(T),rows ,indexes)) = 1; %// Ones at the desired row/column combinations

推荐答案

使用sub2ind解决此问题.您的y确定要使用的列， 行总是总是简单地增加1.通过使用sub2ind，所需的行-列组合将转换为线性索引，然后可以通过矢量化的方式进行处理.

Use sub2ind for this problem. Your y determines the columns to use, the rows are simply always incremented by 1. By using sub2ind the desired row-column-combinations get transformed to a linear index and can then all be adressed in a vectorized way.

y = [3 1 5 3 4 2]; %// column indx
rows = 1:length(y); %// row indx

M = zeros(length(y), max(y)); %// A matrix full of zeros
M(sub2ind(size(M),rows ,y)) = 1; %// Ones at the desired row/column combinations

这篇关于如何从标签向量生成标签矩阵?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！