Hidden Markov models (HMM) are a widelyused tool for sequence modelling. In the sequence classification case, thestandard approach consists of training one HMM for each class and then using a standard Bayesian classification rule. In thispaper, we introduce a novel classification scheme for sequences based on HMMs, which is obtained bye xtending the recentlyproposed similarity-based classification paradigm to HMM-based classification. In this approach, each object is described bythe vector of its similarities with respect to a predetermined set of other objects, where these similarities are supported byHMMs. A central problem is the high dimensionalityof resulting space, and, to deal with it, three alternatives are investigated.Synthetic and real experiments show that the similarity-based approach outperforms standard HMM classification schemes.
Similarity-based classification of sequences using Hidden Markov Models
BICEGO, Manuele;MURINO, Vittorio;
2004-01-01
Abstract
Hidden Markov models (HMM) are a widelyused tool for sequence modelling. In the sequence classification case, thestandard approach consists of training one HMM for each class and then using a standard Bayesian classification rule. In thispaper, we introduce a novel classification scheme for sequences based on HMMs, which is obtained bye xtending the recentlyproposed similarity-based classification paradigm to HMM-based classification. In this approach, each object is described bythe vector of its similarities with respect to a predetermined set of other objects, where these similarities are supported byHMMs. A central problem is the high dimensionalityof resulting space, and, to deal with it, three alternatives are investigated.Synthetic and real experiments show that the similarity-based approach outperforms standard HMM classification schemes.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
