TABLE 4.

Classification model performance (calculated with LOO cross-validation) for models trained with features from Feature Set 2.

	MARS		RFs		GPR
	FS2-MARS	FS2-PCA	FS2-RF	FS2-PCA	FS2-GPR	FS2-PCA
Accuracy	0.62	0.67	0.84	0.67	0.69	0.67
Precision	0.68	0.68	0.82	0.73	0.70	0.67
Recall	0.82	0.93	0.96	0.79	0.93	1.0
MCC	0.04	0.11	0.62	0.22	0.23	0

Bold values indicate performance metrics for recommended, best-performing models.