For a data structure NET, the matrix of vectors X is input into the Support Vector Machine described by NET and the matrix of outputs Y is computed. NET must have non-empty fields NET.sv, NET.svcoeff and NET.bias, these fields are set during training by SVMTRAIN.
X must contain one input vector per row.
Y is a column vector with one entry for each input vector in X.
Y(i) is the SVM output for input vector X(i,:), it is
+1, if X(i,:) is classified as belonging to class 1
-1, if X(i,:) is classified as belonging to class -1