Modified AIC rule for model selection in combination with prior estimated noise models.
Johan SchoukensYves RolainRik PintelonPublished in: Autom. (2002)
Keyphrases
- model selection
- cross validation
- parameter estimation
- hyperparameters
- bayesian model averaging
- bayesian methods
- model selection criteria
- information criterion
- regression model
- statistical inference
- bayesian learning
- error estimation
- sample size
- mixture model
- gaussian process
- marginal likelihood
- statistical learning
- bayesian approaches
- feature selection
- selection criterion
- variable selection
- leave one out cross validation
- closed form
- machine learning
- posterior distribution
- generalization bounds
- missing data
- prior knowledge
- hypothesis tests
- training data
- automatic model selection