function X_rob = robust_norm(X)
% ROBUST_NORM - returns the "robust" normalization of the data in X
%   
% Input: 
%   X: [n,p] = number of observations by number of dimensions 
% 

[n,p] = size(X); 

m = median(X,1); 
s = iqr(X,1); 

X_rob = (X - repmat(m,n,1))./repmat(s,n,1);