The Necessity of Leave One Subject Out (LOSO) Cross Validation for EEG Disease Diagnosis.
Sajeev KunjanTyler S. GrummettKenneth J. PopeDavid M. W. PowersSean P. FitzgibbonT. BastiampillaiM. BattersbyTrent W. LewisPublished in: BI (2021)
Keyphrases
- cross validation
- disease diagnosis
- model selection
- support vector
- variable selection
- training set
- generalization error
- comparative analysis
- hyperparameters
- cross validated
- nearest neighbor classifiers
- unseen data
- binary classifiers
- nearest neighbor
- image processing
- brain computer interface
- ls svm
- information criterion
- learning algorithm
- data sets
- database