3 %> @brief
"Softmax" Transformation
5 %> Returns a matrix where each row sums to one and the elements are exponentially proportional to their respective originals
8 %> Kuncheva, Combining Pattern Classifiers, 2004, section 5.1, page 152
14 Y = Xe./(realmin+repmat(sum(Xe, 2), 1, size(Xe, 2)));
Least-squares classifier.