A variable step learning algorithm for Gaussian mixture models based on the Bhattacharyya coefficient and correlation coefficient criterion.
Weishi PengYangwang FangRenjun ZhanPublished in: Neurocomputing (2017)
Keyphrases
- correlation coefficient
- gaussian mixture model
- learning algorithm
- mixture model
- bhattacharyya coefficient
- em algorithm
- speaker recognition
- feature vectors
- probability density function
- maximum likelihood
- feature space
- expectation maximization
- maximum likelihood criterion
- standard deviation
- training data
- root mean square error
- speaker identification
- gaussian mixture
- machine learning
- chi square
- covariance matrices
- bayesian information criterion
- principal components
- gaussian distribution
- supervised learning
- active learning
- pattern recognition
- root mean squared error
- density estimation
- computer vision
- probability density
- mean shift
- probabilistic model
- high dimensional
- support vector
- stepwise regression
- reinforcement learning