Subspace Gaussian Mixture Models for vectorial HMM-states representation.
Mohamed BouallegueDriss MatroufMickael RouvierGeorges LinarèsPublished in: ASRU (2011)
Keyphrases
- gaussian mixture model
- mixture model
- covariance matrices
- hidden markov models
- feature space
- feature vectors
- speaker recognition
- probability density
- maximum likelihood
- expectation maximization
- finite mixtures
- principal component analysis
- speaker identification
- density estimation
- low dimensional
- variational bayes
- supervised learning
- gaussian distribution
- model selection
- probability density function
- speech recognition
- unsupervised learning
- training data
- text classification
- language model
- text mining
- pattern recognition