Reliability of Cross-Validation for SVMs in High-Dimensional, Low Sample Size Scenarios.
Sascha KlementAmir Madany MamloukThomas MartinetzPublished in: ICANN (1) (2008)
Keyphrases
- confidence intervals
- sample size
- cross validation
- model selection
- support vector
- high dimension
- small sample size
- small sample
- high dimensional
- variable selection
- feature selection
- regression problems
- hyperparameters
- unseen data
- ls svm
- generalization error
- kernel function
- machine learning
- covariance matrix
- high dimensionality
- random sampling
- rbf kernel
- regression model
- information criterion
- feature space
- upper bound
- logistic regression
- dimensionality reduction
- training set
- support vector machine
- multi class
- progressive sampling
- vc dimension
- generalization ability
- microarray data
- kernel methods
- svm classifier
- high dimensional data
- worst case
- nearest neighbor
- input space
- dimension reduction
- data points
- optimality conditions
- objective function
- data sets
- low dimensional
- support vector machine svm