Table 2.

Performance metrics evaluation of predictive models in LACC and LALC testing sets using the three ML pipelines.

ML	Data	BAcc(%)		MCC [−1,+1]
		LACC	LALC	LACC	LALC
MR	Untransformed	60	27	0.19	−0.39
	ComBat	68	69	0.35	0.31
	B-ComBat	70	71	0.37	0.31
	M-ComBat	68	69	0.35	0.31
	BM-ComBat	70	71	0.35	0.31
RF	Untransformed	79	82	0.69	−0.15
	ComBat	84	83	0.72	0.18
	B-ComBat	86	85	0.73	0.23
	M-ComBat	85	84	0.70	0.16
	BM-ComBat	89	86	0.89	0.59
SVM	Untransformed	75	41	0.53	−0.13
	ComBat	79	48	0.55	−0.08
	B-ComBat	82	62	0.63	0.18
	M-ComBat	79	58	0.55	0.12
	BM-ComBat	83	60	0.67	0.17