Continuous HMM with state memberships provided by Takagi-Sugeno fuzzy rule systems (TSFRS)

Author

Popescu, Mihail ; Gader, Paul

Author_Institution

Missouri Univ., Columbia, MO, USA

fYear

2002

fDate

2002

Firstpage

167

Lastpage

171

Abstract

In this paper we develop an EM based training algorithm for a Takagi-Sugeno fuzzy rule system (TSFRS). Since the training is unsupervised, no target values are needed. The TSFRS models the degree of membership based on a given distribution that can be modified by changing the consequence of the rules or by rule pruning. We use this training algorithm to train a hidden Markov model (HMM) with state memberships provided by TSFRS using a modified Baum-Welch algorithm. This representation has the advantage of being transparent, since one can analyze and modify the rules that form the membership TSFRS.

Keywords

fuzzy set theory; fuzzy systems; hidden Markov models; knowledge based systems; unsupervised learning; Baum-Welch algorithm; EM algorithm; Takagi-Sugeno fuzzy rule system; expectation maximization algorithm; hidden Markov model; rule pruning; training algorithm; unsupervised learning; Fuzzy sets; Fuzzy systems; Gaussian processes; Handwriting recognition; Hidden Markov models; Input variables; Neural networks; Process design; Speech recognition; Takagi-Sugeno model;

fLanguage

English

Publisher

ieee

Conference_Titel

Fuzzy Information Processing Society, 2002. Proceedings. NAFIPS. 2002 Annual Meeting of the North American

Print_ISBN

0-7803-7461-4

Type

conf

DOI

10.1109/NAFIPS.2002.1018049

Filename

1018049