How to predict responses of new data from a crossvalidated SVR model

Hi there

I have trained and cross validated my Support Vector Machine regressor model (CValidated_Mdl) with KFold cross validation technique.

I know I can predict responses by using YFit= kfoldPredict(CValidated_Mdl) where YFit are the new responses predicted by the model.

I also have a new set of data(unseen by model) which I will like to use to test the performance of my CValidated_Mdl.

This new and unseen data is called X_test.

I am not sure of how to use the cross validated model (CValidated_Mdl) to predict responses from the X_test data

I have tried YFit= kfoldPredict(CValidated_Mdl, X_test) without success

Can you advise please

Thank You

Answer 1

[Dr. JANAK TRIVEDI]

You can use the predict function in MATLAB to predict responses using the cross-validated model (CValidated_Mdl) and the new data X_test. The code would look like this:

YFit = predict(CValidated_Mdl.Trained{1}, X_test);

Note that CValidated_Mdl.Trained{1} is the trained SVM model for the first fold in the cross-validation, you can use any fold that you think has the best performance.