In this research, the Fuzzy-Neural Network (fuzzy-NN) model was proposed for Speaker-Independent Thai polysyllabic word recognition. Various fuzzy membership functions on linguistic properties were used to convert exact features extracted from input speech to the fuzzy membership values. The fuzzy membership values were arranged to be new input vector of Multilayer Perceptron (MLP) neural network. The binary desired outputs were used during training. 70 Thai words consist of ten numerals, the others were single-syllable, double-syllable and triple-syllable, 20 words in each group, were used for system evaluation. In order to improve recognition accuracy, number of syllable and tonal level detected were conducted for speech preclassification. The Pi fuzzy membership function provided the best recognition accuracy among other functions; Trapezoidal, and Triangular function. Under an optimal condition, the achieved recognition error rates were 5.6% on dependent test and 6.7% on independent test, which were respectively 3.3% and 3.4% decreasing from the conventional Neural Network system.
Cite as: Wutiwiwatchai, C., Jitapunkul, S., Ahkuputra, V., Maneenoi, E., Luksaneeyanawin, S. (1998) Thai polysyllabic word recognition using fuzzy-neural network. Proc. 5th International Conference on Spoken Language Processing (ICSLP 1998), paper 0350, doi: 10.21437/ICSLP.1998-410
@inproceedings{wutiwiwatchai98b_icslp,
author={Chai Wutiwiwatchai and Somchai Jitapunkul and Visarut Ahkuputra and Ekkarit Maneenoi and Sudaporn Luksaneeyanawin},
title={{Thai polysyllabic word recognition using fuzzy-neural network}},
year=1998,
booktitle={Proc. 5th International Conference on Spoken Language Processing (ICSLP 1998)},
pages={paper 0350},
doi={10.21437/ICSLP.1998-410}
}