Model Selection for the ℓ2-SVM by Following the Regularization Path.
Rémi BonidalSamy TindelYann GuermeurPublished in: Trans. Comput. Collect. Intell. (2014)
Keyphrases
- model selection
- feature selection
- selection criterion
- automatic model selection
- support vector
- cross validation
- parameter selection
- support vector machine svm
- learning machines
- hyperparameters
- bayesian learning
- machine learning
- statistical learning
- kernel machines
- support vector machine
- error estimation
- grid search
- regression model
- sample size
- svm classifier
- statistical learning theory
- parameter estimation
- gaussian process
- mixture model
- kernel function
- training data
- variable selection
- generalization ability
- model selection criteria
- information criterion
- generalization error
- support vectors
- feature vectors
- statistical inference
- prior information
- marginal likelihood
- knn
- kernel matrix
- selection criteria
- data dependent
- feature space
- bayesian model selection
- training set