similarity_pearson(data)

function R = similarity_pearson(data)
% pearson coefficients between every two columns 
% input matrix: data --- nrow rows * ncol columns
% output matrix:   R --- ncol * ncol matrix

[nrow,ncol] = size(data);
x = mean(data);
data = data-repmat(x,nrow,1);
R = ones(ncol,ncol);

for i = 1:ncol-1
 x = data(:,i);
 X = sqrt(x'*x);
 for j = i+1:ncol
   y = data(:,j);
   xy = x'*y;
   Y = sqrt(y'*y);
   XY = X*Y;
   if XY > 0
     sim = xy/(XY);
     R(i,j) = sim;
     R(j,i) = sim; 
   end
 end
end