A blocked Gibbs sampler for NGG-mixture models via a priori truncation.
Raffaele ArgientoIlaria BianchiniAlessandra GuglielmiPublished in: Stat. Comput. (2016)
Keyphrases
- mixture model
- gibbs sampler
- gibbs sampling
- bayesian inference
- em algorithm
- probabilistic model
- random fields
- expectation maximization
- generative model
- posterior distribution
- markov chain
- gaussian mixture model
- parameter estimation
- maximum likelihood
- model selection
- unsupervised learning
- language model
- probability density function
- hyperparameters
- maximum a posteriori
- bayesian framework
- probability distribution
- approximate inference
- machine learning
- text classification
- cross validation
- latent variables
- least squares
- feature vectors
- bayesian networks
- training data