sampledatafromclass.m.svn-base

a function inside machine learning

function [newX, newY] = sampleDataFromClass(X, y, sampleSize, class); 
%Take a random subset of the data, but ensure we have  
numExamples = size(X, 1); 

classIndices = find(y == class); 
numClassMembers = length(classIndices); 
sampleSize = min(sampleSize, numClassMembers); 

permutationVector = randperm(numClassMembers);
sampleIndices = classIndices(permutationVector(1:sampleSize));

newX = X(sampleIndices, :); 
newY = y(sampleIndices, :); 

fprintf('Subsampled data from %d to %d examples, from label %d\n', numClassMembers, sampleSize, class);