Boosting long-term adaptation of hidden-Markov-models: incremental splitting of probability density functions.
Udo BubHarald HögePublished in: ICASSP (1998)
Keyphrases
- hidden markov models
- probability density function
- density function
- density estimation
- probability distribution
- kernel density estimation
- speech recognition
- gaussian mixture model
- machine learning
- posterior probability
- mixture model
- learning algorithm
- expectation maximization
- feature selection
- gaussian mixture
- sequential data
- em algorithm
- baum welch
- conditional random fields
- statistical methods
- bayesian framework
- distance measure
- probabilistic model
- feature vectors
- reinforcement learning
- computer vision