Just use the following.
[cm, order] = confusionmat(real_values,predicted_values)
Assuming your confusionmatrix is 2x2, it will look like this:
%CONFUSIONMAT OUTPUT
%%%%%%%%%%%%%%%%%%%%%%
% predicted
% | 0 1
% r -----------
% e 0 | TN FP
% a 1 | FN TP
% l
%%%%%%%%%%%%%%%%%%%%%%
*EDIT
confusionmat() does not calculate sensitivity, accuracy, F-score and those kind of derivations for you, you need to do them yourself. Just check this out, it's pretty easy.
