Hybrid approach of selecting hyperparameters of support vector machine for regression.
Jin-Tsong JengPublished in: IEEE Trans. Syst. Man Cybern. Part B (2006)
Keyphrases
- hyperparameters
- support vector
- support vector machine
- model selection
- regression problems
- regression function
- gaussian processes
- cross validation
- grid search
- support vector regression
- gaussian process
- bayesian inference
- svm classifier
- bayesian framework
- feature selection
- regression model
- closed form
- prior information
- em algorithm
- random sampling
- noise level
- kernel function
- covariance function
- generalization ability
- maximum a posteriori
- support vector machine svm
- decision function
- parameter optimization
- machine learning
- incremental learning
- maximum likelihood
- gaussian process regression
- incomplete data
- prior knowledge
- sample size
- missing values
- feature vectors
- edge detection
- parameter settings