TABLE 4.

Classification results of disease fatigue against non-disease fatigue over four datasets using four classifiers.

Classifiers	Data sets	Sensitivity (%)	Specificity (%)	F1	Precision (%)	Accuracy (%)	AUC
Logistic regression	Tongue	60.82 ± 4.26	64.49 ± 5.71	0.6192 ± 0.0272	63.32 ± 3.41	62.65 ± 2.67	0.6666 ± 0.0284
	Pulse	62.86 ± 6.88	63.67 ± 4.36	0.6297 ± 0.0476	63.33 ± 3.37	63.27 ± 3.93	0.6990 ± 0.0370
	Tongue & Pulse	67.76 ± 6.11	67.14 ± 6.35	0.6749 ± 0.0457	67.49 ± 4.75	67.45 ± 4.32	0.7395 ± 0.0415
	Tongue & Pulse & Age & BMI	84.90 ± 3.56	86.12 ± 4.36	0.8542 ± 0.0181	86.18 ± 3.67	85.51 ± 1.87	0.9160 ± 0.0136
SVM	Tongue	56.33 ± 4.58	68.78 ± 7.13	0.6004 ± 0.0303	64.71 ± 4.70	62.55 ± 3.10	0.6470 ± 0.0430
	Pulse	64.29 ± 4.76	65.71 ± 3.00	0.6468 ± 0.0313	65.21 ± 2.20	65.00 ± 2.50	0.7035 ± 0.0243
	Tongue & Pulse	65.10 ± 6.42	68.57 ± 5.34	0.6617 ± 0.0506	67.48 ± 4.60	66.84 ± 4.59	0.7203 ± 0.0389
	Tongue & Pulse & Age & BMI	85.31 ± 5.83	82.24 ± 5.56	0.8399 ± 0.0439	82.92 ± 4.78	83.78 ± 4.39	0.9106 ± 0.0365
Random forest	Tongue	61.84 ± 6.89	68.37 ± 4.00	0.6380 ± 0.0521	66.08 ± 4.14	65.10 ± 4.40	0.6803 ± 0.0630
	Pulse	60.61 ± 5.78	61.84 ± 6.89	0.6097 ± 0.0498	61.52 ± 5.39	61.22 ± 5.14	0.6582 ± 0.0509
	Tongue & Pulse	66.94 ± 5.38	70.41 ± 5.79	0.6806 ± 0.0356	69.57 ± 4.22	68.67 ± 3.32	0.7423 ± 0.0444
	Tongue & Pulse & Age & BMI	84.90 ± 4.40	81.63 ± 4.74	0.8353 ± 0.0344	82.33 ± 3.99	83.27 ± 3.48	0.8959 ± 0.0254
Neural network	Tongue	62.45 ± 5.86	63.88 ± 5.78	0.6281 ± 0.0403	63.47 ± 3.72	63.16 ± 3.49	0.6639 ± 0.0255
	Pulse	65.31 ± 7.36	64.90 ± 6.04	0.6500 ± 0.0390	65.17 ± 2.89	65.10 ± 2.77	0.7087 ± 0.0330
	Tongue & Pulse	65.7 ± 12.27	70.2 ± 13.20	0.6664 ± 0.0633	70.34 ± 6.27	67.96 ± 3.33	0.7454 ± 0.0349
	Tongue & Pulse & Age & BMI	85.31 ± 6.57	86.33 ± 1.84	0.8562 ± 0.0353	86.19 ± 1.50	85.82 ± 3.01	0.9239 ± 0.0174