On the use of marginal posteriors in marginal likelihood estimation via importance sampling.
Konstantinos PerrakisIoannis NtzoufrasEfthymios G. TsionasPublished in: Comput. Stat. Data Anal. (2014)
Keyphrases
- importance sampling
- marginal likelihood
- approximate inference
- posterior distribution
- monte carlo
- hyperparameters
- model selection
- gaussian process
- graphical models
- parameter estimation
- probability distribution
- particle filter
- probabilistic inference
- markov chain monte carlo
- kalman filter
- markov chain
- belief propagation
- closed form
- cross validation
- particle filtering
- bayesian networks
- message passing
- bayesian framework
- random sampling
- bayesian inference
- latent variables
- em algorithm
- maximum likelihood
- support vector
- noise level
- incremental learning
- prior information