centersparsedata.m.svn-base

a function inside machine learning

function cX = centerSparseData(X) 
%Centers a sparse matrix by taking the mean of each column (feature) of M and 
%subtracting the mean from feature value. NOTE: This often makes it non
%sparse!
%
%inputs 
% M is the matrix to be centered
%
%outputs 
% cX is the centered matrix 

[numExamples, numFeatures] = size(X); 
cX = sparse(numExamples, numFeatures); 

meanX = mean(X);

for i=1:numFeatures 
    cX(:, i) = X(:, i) - meanX(i); 
end