how to compute correlation coefficient without using any matlab functions Except the mean function?

how i can compute a correlation coefficient which accepts as input two arrays x and y of the same length representing the data sample and, returns their correlation coefficient (i loaded the x and y inputs as arrays). without using any matlab functions Except the mean function.

Answer 1

[Star Strider]

Using the Wikipedia definition of the Pearson product-moment correlation for a sample, this works:

x = randi(10,25,1);                         % Create Data
y = randi(10,25,1);                         % Create Data

xm = mean(x);
ym = mean(y);

[sxy,s2x,s2y] = deal(0,0,0);
for k1 = 1:length(x)
    sxy = sxy + (x(k1)-xm)*(y(k1)-ym);
    s2x = s2x + (x(k1)-xm)^2;
    s2y = s2y + (y(k1)-ym)^2;
end
r = sxy / (sqrt(s2x)*sqrt(s2y));