Exploiting Generative Models in Discriminative Classiiers

Generative probability models such as hidden Markov models provide a principled way of treating missing information and dealing with variable length sequences. On the other hand, discriminative methods such as support vector machines enable us to construct exible decision boundaries and often result in classiication performance superior to that of the model based approaches. An ideal classiier should combine these two complementary approaches. In this paper, we develop a natural way of achieving this combination by deriving kernel functions for use in discriminative methods such as support vector machines from generative probability models. We provide a theoretical justiication for this combination as well as demonstrate a substantial improvement in the classiication performance in the context of DNA and protein sequence analysis.