# SVM for more than 2 classes

I need to know how to carry out Support Vector Machines (SVM) with more than two classes. Is there a book or reference about this?

You can apply any binary classifier to multiclass classification by using one-versus-all classification. If you have $n$ classes, you would train $n$ classifiers, one for each class. At decision time, you would predict the label given the highest confidence by one of your classifiers. There are other methods as well, but this is the simplest.