How can i create a SVM in the case of multiclass classification?

Hi,
I wrote a code for the identification of digits, now i need to create a svm classifier, but in my case i have more than two classes, in fact i have 10. I have to use the one vs one approach, and if i understood the method correctly i need to train N(N-1)/2 (in my case 45) classifiers to discriminate each pair and then tale the classifier which got more votes. How can i do that? Thanks

Tag

Richiesto:

il 11 Feb 2021

Risposto:

il 14 Feb 2021

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by