https://se.mathworks.com/help/deeplearning/ref/plotconfusion.html

For Test Data Set you have arrays Real_Labels and Predicted_Labels. First, you should calculate amount of TP, TN, FP and FN (i.e. Confusion Matrix) for each class I :

#TP = length (find (Real_Labels = I & Predicted_Labels = I )),

#TN = length (find (Real_Labels =/= I & Predicted_Labels =/= I )),

#FP = length (find (Real_Labels =/= I & Predicted_Labels = I )),

#FN = length (find (Real_Labels = I & Predicted_Labels =/= I )),

After this you can calculate for this class I the Accuracy, Precision, Sensitivity, Specifity , etc. based on these values.

Hi Bhagyasree Kanuparthi ,

If you are using cross validation, then you need to define class performance as follows.

cp = classperf(Label);

pred1 = predict(Mdl,data(test,:));

where Mdl is your classifier model. Test the performance as follows.

cp = classperf(cp, pred1, test);

Acc=cp.CorrectRate; (CorrectRate is accuracy)

Sens=cp.sensitivity;

spec=cp.specificity;

cp.CountingMatrix

OR you may refer the following links.

https://in.mathworks.com/help/deeplearning/ref/confusion.html

https://in.mathworks.com/help/stats/crossval.html