An optimization-based algorithm for model selection using an approximation of Akaike's Information Criterion.
Rodrigo CarvajalGabriel UrrutiaJuan C. AgüeroPublished in: AuCC (2016)
Keyphrases
- information criterion
- model selection
- automatic model selection
- marginal likelihood
- probability model
- cross validation
- bayesian information criterion
- leave one out cross validation
- selection criterion
- mixture model
- parameter estimation
- hyperparameters
- learning algorithm
- machine learning
- sample size
- model selection criteria
- generalization error
- variable selection
- closed form
- free energy
- expectation maximization
- regression model
- maximum likelihood
- objective function
- bayesian framework
- gaussian process
- factor analysis
- widely applicable