Table 2.

Results of the prediction models

Audio_format	Gender	ML model	Accuracy	Ov.precision	Precision_0	Precision_1	Ov.recall	Recall_0	Recall_1	Ov.f1score	f1-score_0	f1-score_1	Weighted AUC
3gp (Android)	Female	LR	0.77	0.77	0.81	0.73	0.77	0.76	0.77	0.77	0.78	0.75	0.85
		KNN	0.72	0.73	0.7	0.77	0.72	0.87	0.55	0.72	0.78	0.64	0.76
		SVM	0.8	0.8	0.8	0.79	0.8	0.84	0.74	0.8	0.82	0.77	0.86
		VC	0.78	0.78	0.81	0.75	0.78	0.79	0.77	0.78	0.8	0.76	0.86
	Male	LR	0.78	0.79	0.87	0.5	0.78	0.85	0.53	0.79	0.86	0.52	0.81
		KNN	0.83	0.83	0.83	0.8	0.83	0.98	0.27	0.79	0.9	0.4	0.84
		SVM	0.84	0.83	0.88	0.67	0.84	0.93	0.53	0.83	0.9	0.59	0.82
		VC	0.84	0.84	0.89	0.64	0.84	0.91	0.6	0.84	0.9	0.62	0.82
m4a (iOS)	Female	LR	0.72	0.72	0.8	0.56	0.72	0.77	0.61	0.72	0.79	0.58	0.75
		KNN	0.68	0.65	0.72	0.5	0.68	0.86	0.29	0.65	0.78	0.37	0.67
		SVM	0.79	0.79	0.81	0.75	0.79	0.91	0.55	0.79	0.86	0.64	0.79
		VC	0.77	0.76	0.8	0.69	0.77	0.89	0.53	0.76	0.84	0.6	0.78
	Male	LR	0.73	0.74	0.83	0.48	0.73	0.8	0.54	0.73	0.81	0.51	0.8
		KNN	0.89	0.89	0.89	0.89	0.89	0.97	0.65	0.88	0.93	0.76	0.81
		SVM	0.85	0.84	0.86	0.76	0.85	0.95	0.58	0.84	0.9	0.67	0.85
		VC	0.89	0.89	0.89	0.89	0.89	0.97	0.65	0.88	0.93	0.76	0.85

The selected models were selected using Recall_1 and weighted AUC and are highlighted in bold. Class 0: no fatigue, class 1: fatigue.

AUC, area under the curve; KNN, K-Nearest Neighbuors; LR, logistic regression; Ov, Overall; SVM, support vector machine; VC, voting classifier.