Noise derived information criterion for model selection.
Joaquín Pizarro JunqueraPedro L. GalindoElisa Guerrero VázquezAndrés Yáñez EscolanoPublished in: ESANN (2002)
Keyphrases
- information criterion
- model selection
- cross validation
- model selection criteria
- automatic model selection
- bayesian information criterion
- sample size
- hyperparameters
- marginal likelihood
- generalization error
- machine learning
- probability model
- selection criterion
- variable selection
- parameter estimation
- noise level
- regression model
- mixture model
- gaussian process
- leave one out cross validation
- factor analysis
- kernel matrix
- missing data
- feature selection