A nonlinear least squares quasi-Newton strategy for LP-SVR hyper-parameters selection.
Pablo Rivas-PereaJuan Cota-RuizJosé-Gerardo RosilesPublished in: Int. J. Mach. Learn. Cybern. (2014)
Keyphrases
- hyperparameters
- model selection
- support vector
- cross validation
- support vector regression
- bayesian inference
- em algorithm
- bayesian framework
- random sampling
- sample size
- linear programming
- closed form
- kernel function
- maximum a posteriori
- gaussian process
- maximum likelihood
- regression model
- ls svm
- support vector machine
- optimal solution
- objective function
- image processing
- neural network