A Bayesian Kernel Logistic Discriminant Model: An Improvement to the Kernel Fisher's Discriminant

Riadh Ksantini, Djemel Ziou, Bernard Colin, François Dubeau

Abstract: The Kernel Fisher's Discriminant (KFD) has proven to be competitive to several state-of-the-art classifiers. However, it is assuming equal covariance structure for all transformed classes, which is not true in many applications. In this paper, we propose a novel Bayesian Kernel Logistic Discriminant model (BKLD) which goes one step further by representing each transformed class by its own covariance matrix. This can perform better than the KFD. An extensive comparison of the BKLD to the KFD and to other state-of-the-art non-linear classifiers is performed.