Development and Validation of a Magnetic Resonance Imaging-Based Machine Learning Model for TMJ Pathologies

Table 5

Evaluation for diagnostic performance by four indicators set for TMJ disc: precision, recall, F1-score, and support in the training set. “1” indicates normal. “2” indicates anterior disc displacement.

		Indicators	KNN	SVM	XGBoost	RF	LR	DT
Training	1	Precision	0.84	0.77	0.88	0.999	0.80	0.78
		Recall	0.95	1.00	0.90	0.999	0.93	0.89
		F1-score	0.89	0.87	0.89	0.999	0.86	0.86
		Support	40	40	40	40	40	40
	2	Precision	0.71	0.00	0.64	1.00	0.50	0.54
		Recall	0.42	0.00	0.58	1.00	0.25	0.28
		F1-score	0.53	0.00	0.61	1.00	0.33	0.48
		Support	12	12	12	12	12	12
Testing	1	Precision	0.77	0.79	0.85	0.79	0.79	0.73
		Recall	0.91	1.00	1.00	1.00	1.00	0.81
		F1-score	0.83	0.88	0.92	0.88	0.88	0.77
		Support	11	11	11	11	11	11
	2	Precision	0.00	0.00	1.00	0.00	0.00	0.50
		Recall	0.00	0.00	0.33	0.00	0.00	0.33
		F1-score	0.00	0.00	0.50	0.00	0.00	0.40
		Support	3	3	3	3	3	3