Generalization of the minimum classification error (MCE) training based on maximizing generalized posterior probability (GPP).
Qiang FuAntonio Moreno-DanielBiing-Hwang JuangJian-Lai ZhouFrank K. SoongPublished in: INTERSPEECH (2006)
Keyphrases
- minimum classification error
- posterior probability
- discriminative training
- bayesian framework
- bayesian networks
- generative model
- probability distribution
- statistical pattern recognition
- hidden markov models
- maximum likelihood estimation
- gaussian mixture
- probability density function
- posterior distribution
- conditional probabilities
- probabilistic model
- markov chain monte carlo
- graphical models