Discriminative training of hidden Markov models using a classification measure criterion.
Cristina ChestaAldo GirardiPietro LafaceMario NigraPublished in: ICASSP (1998)
Keyphrases
- discriminative training
- hidden markov models
- minimum classification error
- maximum likelihood
- probabilistic interpretation
- automatic speech recognition
- sequential data
- speech recognition
- sequence classification
- support vector machine
- posterior probability
- markov models
- similarity measure
- co training
- conditional random fields
- maximum likelihood estimation
- machine learning
- training set
- feature space
- pattern recognition
- reinforcement learning
- baum welch
- feature selection