Hi,
You can use the MATLAB function "kfoldPredict" to classify observations in cross-validated classification mode. You can also use MATLAB's built in function "confusionmat" to compute confusion matrix for classification problem. Here is an example as mentioned in the following documentation on how to achieve the same: https://www.mathworks.com/help/stats/confusionmat.html
g1 = [3 2 2 3 1 1]'; % Known groups
g2 = [4 2 3 NaN 1 1]'; % Predicted groups
C = confusionmat(g1,g2) ; % Returns the confusion matrix
In order to evaluate your model's performance, you can use MATLAB's function "perfcurve". Please refer to the following documentation of "perfcurve" for a better understanding: https://www.mathworks.com/help/stats/perfcurve.html
I hope this helps!
