xiaoyili / Sparse_coding

UFLDL

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

关于sparseCodingFeatureCost.m第52行,为何不是UFLDL上的s乘以s的转置,而是s.^2 ?

sinb opened this issue · comments

commented

第52行,
sparsityMatrix = sqrt(groupMatrix_(featureMatrix.^2)+epsilon);
如果按照UFLDL上的公式,不应该是这样吗
sparsityMatrix = sqrt(groupMatrix_(featureMatrix * featureMatrix')+epsilon);
我看网上所有的UFLDL里稀疏编码的matlab code都是这样写的,求教为什么,多谢!