performance - Remove for loop from clustering algorithm in MATLAB -

i trying improve performance of optics clustering algorithm. implementation i've found in open source makes use of loop each sample , can run hours...

i believe use of repmat() function may aid in improving performance when system has enough amount of ram. more welcome suggest other ways of improving implementation.

here code:

x data: [mxn] array m sample size , n feature dimensionality, of time greater one.

[m,n] = size(x);  = 1:m     d(i,:) = sum(((repmat(x(i,:),m,1)-x).^2),2).'; end 

many thanks.

with enough ram play with, can use few approaches here.

approach #1: bsxfun & permute -

d = squeeze(sum(bsxfun(@minus,permute(x,[3 2 1]),x).^2,2)) 

approach #2: pdist & squareform -

d = squareform(pdist(x).^2) 

approach #3 matrix-multiplication based euclidean distance calculations -

xt = x.';  %//' [m,n] = size(x); d = [x.^2 ones(size(x)) -2*x ]*[ones(size(xt)) ; xt.^2 ; xt]; d(1:m+1:end) = 0; 

for performance, bet on approach #3!