Bandwidth Selection for Nadaraya-Watson Kernel Estimator Using Cross-Validation Based on Different Penalty Functions.
Yumin ZhangPublished in: ICMLC (CCIS volume) (2014)
Keyphrases
- cross validation
- nadaraya watson
- kernel regression
- penalty functions
- model selection
- regression model
- random fields
- support vector
- variable selection
- penalty function
- training set
- hyperparameters
- constrained optimization
- regression methods
- parameter estimation
- linear regression
- error estimates
- generalization error
- objective function
- parameter optimization
- markov random field
- higher order
- estimation problems
- genetic algorithm
- machine learning
- least squares
- image processing