Parameter-based reduction of Gaussian mixture models with a variational-Bayes approach.
Pierrick BruneauMarc GelgonFabien PicarougnePublished in: ICPR (2008)
Keyphrases
- variational bayes
- gaussian mixture model
- mixture model
- feature vectors
- em algorithm
- bayesian inference
- maximum likelihood
- latent variables
- expectation maximization
- bayesian learning
- free energy
- gaussian distribution
- feature space
- hyperparameters
- density estimation
- latent dirichlet allocation
- parameter settings
- noise level
- incremental learning
- probability density function
- parameter space
- hidden markov models
- probabilistic model