MATLAB: How to use big matrix for input of Kmeans

kmeansStatistics and Machine Learning Toolbox

Hi
I have a matrix that is 90,000 * 6,000 . it's 'Y'. I need to use it in Kmeans matlab code , But it's too big and computer was crashed, how can i use it ??
the part of my code is here:
nC = 3 ;
opts = statset('Display','final');
X = Y' ;
[idx,ctrs] = kmeans(X,nC,...
'Distance','cosine',...
'Replicates',5,...
'Options',opts);
figure()
silhouette(X,idx,'sqeuclid');
title('K-Means')
How can I use Y for Kmeans?? The Ram is 6 GB.

Best Answer

How many independent variables do you have? I'm used to seeing kmeans used on something like a 2D scatter plot (y vs. x) where you have an N by 2 list of (x,y) coordinates. It seems like you have 6 thousand independent variables instead of 2. What exactly is your Y', which you renamed to X?