Generalization performance of multi-class discriminant models
Abstract
Starting from a direct definition of the notion of margin in the multiclass case , we study the generalization performance of multiclass discriminant systems. In the framework of statistical learning theory, we establish on this performance a bound based on covering numbers. An application to a linear ensemble method w hich estimates the class posterior probabilities provides us with a way to compa re this bound and another one based on combinatorial dimensions, with respect to the capacity measure they incorporate. Experimental results highlight their use fulness for a real-world problem.