Sixth International Conference on Spoken Language Processing
(ICSLP 2000)

Beijing, China
October 16-20, 2000

Fuzzy Entropy Hidden Markov Models for Speech Recognition

Dat Tran, Michael Wagner

Human-Computer Communication Lab., School of Computing, University of Canberra, Australia

A new fuzzy technique called fuzzy entropy (FE) clustering is proposed and applied to hidden Markov models (HMMs) for speech recognition. FE-HMMs, both discrete and con- tinuous, are proposed in this paper. Experimental results in speech recognition show good results for FE models com- pared with fuzzy C-means and conventional models.


Full Paper

Bibliographic reference.  Tran, Dat / Wagner, Michael (2000): "Fuzzy entropy hidden Markov models for speech recognition", In ICSLP-2000, vol.1, 421-424.