Tuning complexity in regularized kernel-based regression and linear system identification: The robustness of the marginal likelihood estimator.
Gianluigi PillonettoAlessandro ChiusoPublished in: Autom. (2015)
Keyphrases
- marginal likelihood
- regression function
- hyperparameters
- model selection
- least squares
- closed form
- support vector
- gaussian process
- sparse approximation
- cross validation
- simple linear
- bayesian inference
- random sampling
- regression model
- gaussian processes
- selection criterion
- sample size
- maximum a posteriori
- prior information
- maximum likelihood
- objective function
- posterior distribution
- parameter estimation
- information criterion
- semi supervised