Key result
Leave-one-subject-out cross-validation for microsleep classification achieved average accuracies of 90.2% without balancing and 89.1% with between-subjects balancing.
Why the study?
Inter-individual differences in EEG feature distributions during microsleep raise questions regarding the generalizability and validity of classification methodology for future subjects.
Sample balancing might not be essential for microsleep classification if leave-one-subject-out cross-validation is performed over a sufficiently large number of subjects.
Answers are grounded in this paper and cite their sources.
Subject-specific EEG variability may limit microsleep detection reliability; leaves open generalizability for future subjects without larger validation.
Inter-individual differences in the feature distribution of electroencephalograms (EEG) during microsleep (MS) raise questions about the generalizability of the methodology, because methodology should have the same validity for subjects to be included in the analysis in the future as for subjects included so far. We address this question using leave-one-subject- out cross-validation (LOSO CV) to simulate inclusion of test data from future subjects. Investigations are based on EEG of 70 subjects across four studies conducted in our driving simulation lab. 9,297 MS and 10,264 counter-examples of sustained attention (SA) were included in the analysis. Three variants of sample balancing are compared: (1) none, (2) overall balancing, and (3) between-subjects balancing. Resulting classification accuracies range from 80% to 100% for 67 subjects, but for 3 subjects they are at 66.5%, 73.1%, 78.3%. Averaged across subjects, accuracies were (1) 90.2%, (2) 90.2% and (3) 89.1%. Thus, balancing might not be essential if LOSO CV is performed over a sufficiently large number of subjects. Results of standard CV must be regarded as optimistically biased.
No takes yet. Share an insight, caveat, or question.
Pauli et al. (2021) studied Microsleep (n=70). Leave-one-subject-out cross-validation (LOSO CV) with sample balancing vs. No balancing was evaluated on Classification accuracy. Leave-one-subject-out cross-validation for microsleep classification achieved average accuracies of 90.2% without balancing and 89.1% with between-subjects balancing.
Synapse has enriched 10 related papers. Consider them for comparative context: