. Author manuscript; available in PMC: 2019 Nov 29.

Published in final edited form as: IEEE Access. 2018 Nov 29;6:77796–77806. doi: 10.1109/ACCESS.2018.2884126

TABLE 5:

Classifiers, features used, and performance statistics for best AUC of diagnostic experiment (experiment 1) using RIDER features

Experiment	Features Set	Classifier	Feature Selector	Features Selected	Performanc Statistics
Experiment	Features Set	Classifier	Feature Selector	Features Selected	AUC (95% CI)	Specificity (95% CI)	Sensitivity (95% CI)
Train C1T1 & Test C2T2	Rider features	Random Forests	None	All Rider Stable features	0.82 (0.79-0.85)	0.92 (0.88-0.96)	0.56 (0.46-0.65)
Train on C1T1 and C1 delta features & Test on C2T2 and C2 delta features (i.e., C2T2 - C2T1)	Rider features	Random Forests	None	All Rider Stable features	0.858 (0.83-0.89)	0.94 (0.91-0.97)	0.65 (0.56-0.74)
p value (p <0.05)					0.381