Density-Preserving Sampling: Robust and Efficient Alternative to Cross-Validation for Error Estimation.
Marcin BudkaBogdan GabrysPublished in: IEEE Trans. Neural Networks Learn. Syst. (2013)
Keyphrases
- nearest neighbor
- cross validation
- error estimation
- model selection
- error estimates
- training set
- generalization error
- nearest neighbor classifiers
- sample size
- hyperparameters
- variable selection
- cross validated
- regression problems
- support vector
- decision trees
- regularization parameter
- machine learning
- random sampling
- data mining
- feature space
- gaussian process