Stepwise Feature Selection by Cross Validation for EEG-based Brain Computer Interface.
Kenji TanakaTakio KuritaFriedrich MeyerLuc BerthouzeTohru KawabePublished in: IJCNN (2006)
Keyphrases
- cross validation
- motor imagery
- feature selection
- model selection
- event related
- selected features
- support vector
- forward selection
- hyperparameters
- brain computer interface
- variable selection
- regression problems
- cross validated
- text categorization
- eeg signals
- regularization parameter
- generalization error
- machine learning
- sample size
- leave one out cross validation
- mutual information
- feature set
- feature extraction
- information criterion
- error estimates
- unseen data
- permutation tests
- training set
- small sample
- feature subset
- eeg data
- nearest neighbor classifiers
- feature space
- support vector machine
- multi class
- meta learning
- data sets
- data mining