SVM Prediction Algorithm-Train

hI,

I am creating a SVM to predict the class of a set of predictor variables. I have seen fitcsvm is the way to train the algorithm. Could you please indicate how to use only 30% of the whole dataset to train the algorithm?

Thanks in advance,

Regards

Answer 1

[Ameer Hamza]

See the cross-validation options on the fitcsvm documentation page: https://www.mathworks.com/help/releases/R2020a/stats/fitcsvm.html#d120e287869