A new hybrid discriminative/generative model using the full-covariance multivariate generalized Gaussian mixture models.
Fatma NajarSami BourouisNizar BouguilaSafya BelghithPublished in: Soft Comput. (2020)
Keyphrases
- generative model
- gaussian mixture model
- mixture model
- discriminative models
- em algorithm
- discriminative learning
- finite mixtures
- generative and discriminative models
- covariance matrices
- expectation maximization
- probabilistic model
- semi supervised
- fisher kernel
- speaker recognition
- probability density function
- density estimation
- bayesian framework
- posterior probability
- gaussian mixture
- maximum likelihood
- gaussian distribution
- prior knowledge
- markov chain monte carlo
- topic models
- latent dirichlet allocation
- covariance matrix
- maximum likelihood estimation
- feature vectors
- likelihood function
- unsupervised learning
- exponential family
- sparse coding
- feature extraction
- hyperparameters
- semi supervised learning
- graphical models
- text mining
- training data