Hidden Markov model estimation based on alpha-EM algorithm: Discrete and continuous alpha-HMMs.
Yasuo MatsuyamaPublished in: IJCNN (2011)
Keyphrases
- hidden markov models
- em algorithm
- baum welch
- maximum likelihood estimation
- parameter estimation
- expectation maximization
- mixture model
- log likelihood function
- maximum likelihood
- density estimation
- gaussian mixture model
- hidden states
- gaussian mixture
- speech recognition
- sequential data
- generative model
- viterbi algorithm
- hidden state
- hyperparameters
- estimate the model parameters
- probability density function
- log likelihood
- maximum a posteriori
- conditional random fields
- markov models
- discrete data
- incomplete data
- bayesian framework
- prior knowledge
- hidden variables
- structure learning
- machine learning
- gibbs sampling
- discriminative training