Classification error from the theoretical Bayes classification risk

  title={Classification error from the theoretical Bayes classification risk},
  author={Erik McDermott and Shigeru Katagiri},
This article shows that the Minimum Classification Error (MCE) criterion function commonly used for discriminative design of speech recognition systems is equivalent to a Parzen window based estimate of the theoretical Bayes classification risk. In this analysis, each training token is mapped to the center of a Parzen kernel in the domain of a suitably defined random variable. The kernels are summed to produce a density estimate; this estimate in turn can easily be integrated over the domain of… CONTINUE READING


Publications referenced by this paper.
Showing 1-6 of 6 references

Pattern Recognition Using a Family of Design Algorithms based upon the Generalized Probabilistic Descent Method

  • S. Katagiri, Juang, B.-H, Lee, C.-H
  • Proceedings of the IEEE,
  • 1998
2 Excerpts

Discriminative Training

  • Juang, B.-H, S. Katagiri
  • Journal of the Acoustical Society of Japan…
  • 1992

Segmental GPD Training of HMM Based Speech Recognizer

  • W. Chou, Juang, B.-H, Lee, C.-H
  • Proceedings of the IEEE, International Conference…
  • 1992
1 Excerpt

New Discriminative Training Algorithms Based on the Generalized Descent Method

  • S. Katagiri, Lee, C.-H, Juang, B.-H
  • Proceedings of the 1991 IEEE Workshop on Neural…
  • 1991
1 Excerpt

Similar Papers

Loading similar papers…