Few paths, fewer words: model selection with automatic structure functions.
Bjørn Kjos-HanssenPublished in: CoRR (2016)
Keyphrases
- model selection
- cross validation
- hyperparameters
- bayesian learning
- statistical inference
- parameter estimation
- regression model
- machine learning
- selection criterion
- generalization error
- error estimation
- statistical learning
- meta learning
- model selection criteria
- sample size
- variable selection
- gaussian process
- marginal likelihood
- feature selection
- mixture model
- bayesian methods
- parameter determination
- leave one out cross validation
- em algorithm
- intelligent data mining
- automatic model selection
- bayesian information criterion
- real world
- image segmentation
- motion segmentation