In this work, we make a contribution to natural speech dialogue act detection. We focus our attention on the dialogue act classification using a Bayesian approach. Our classifier is tested on two corpora, the Switchboard and the Basurde tasks. A combination of a naive Bayes classi- fier and n-grams is used. The impact of different smoothing methods (Laplace and Witten Bell) and n-grams in classification are studied. With respect to the Switchboard corpus, an accuracy of 66% is achieved using a uniform naive Bayes classifier, 3-grams and Laplace smoothing to avoid zero probabilities. For the Basurde corpus, our system achieves performances similar to other methodologies we have previously tested. Through a combination of a naive Bayes classifier with 2-grams and Witten Bell smoothing we achieve the best accuracy of 89%. These results show that a Bayesian approach is well suited for these tasks.
Cite as: Grau, S., Sanchis, E., Castro, M.J., Vilar, D. (2004) Dialogue act classification using a Bayesian approach. Proc. 9th Conference on Speech and Computer (SPECOM 2004), 495-499
@inproceedings{grau04_specom,
author={Sergio Grau and Emilio Sanchis and Maria Jose Castro and David Vilar},
title={{Dialogue act classification using a Bayesian approach}},
year=2004,
booktitle={Proc. 9th Conference on Speech and Computer (SPECOM 2004)},
pages={495--499}
}