Sixth International Conference on Spoken Language Processing
(ICSLP 2000)

Beijing, China
October 16-20, 2000

Efficient Mixed-Order Hidden Markov Model Inference

Ludwig Schwardt, Johan du Preez

Department of Electrical and Electronic Engineering, University of Stellenbosch, Matieland, South Africa

Recent studies have shown that high-order hidden Markov models (HMMs) are feasible and useful for spoken language processing. This paper extends the fixed-order versions to ergodic mixedorder HMMs, which allow the modelling of variable-length contexts with significantly less parameters. A novel training procedure automatically infers the number of states and the topology of the HMM from the training set, based on information-theoretic criteria. This is done by incorporating only high-order contexts with sufficient support in the data. The mixed-order training algorithm is faster than fixed-order methods, with similar classifi- cation performance in language identification tasks.


Full Paper

Bibliographic reference.  Schwardt, Ludwig / Preez, Johan du (2000): "Efficient mixed-order hidden Markov model inference", In ICSLP-2000, vol.2, 238-241.