We present automatic language recognition results using high-order hidden Markov models (HMM) and the recently developed ORder rEDucing (ORED) and Fast Incremental Training (FIT) HMM algorithms. We demonstrate the efficiency and accuracy of pseudo-phoneme context and duration modelling mixed-order HMMs as well as fixed order HMMs over conventional approaches. For a two language problem, we show that a third-order FIT trained HMM gives a test set accuracy of 97.4% compared to 89.7% for a conventionally trained third-order HMM. A first-order model achieved 82.1% accuracy on the same problem.
Cite as: Preez, J.A.d., Weber, D.M. (1998) Automatic language recognition using high-order HMMs. Proc. 5th International Conference on Spoken Language Processing (ICSLP 1998), paper 1074, doi: 10.21437/ICSLP.1998-209
@inproceedings{preez98b_icslp,
author={J. A. du Preez and D. M. Weber},
title={{Automatic language recognition using high-order HMMs}},
year=1998,
booktitle={Proc. 5th International Conference on Spoken Language Processing (ICSLP 1998)},
pages={paper 1074},
doi={10.21437/ICSLP.1998-209}
}