Skip to main content
. 2020 Feb 14;20(4):1029. doi: 10.3390/s20041029

Figure 5.

Figure 5

Graphic representation of Leave-One-Subject-Out Cross-Validation (LOSOCV) for 27 subjects: Oversampling of minority class in training set in each iteration with Synthetic Minority Oversampling Technique (SMOTE); Selection of best feature subset with Correlation-based Feature Subset Selection (CFSS) in training set and adjusting of features in testing set; calculation of average accuracy (A) and F-measure (F) (separately for both classes) across all iterations.