Table 2.

Results of AD vs. NC classification on both ADNI-2 and MIRIAD datasets, with models trained on the ADNI-1 dataset.

	ADNI-2					MIRIAD
	ROI	VBM	CLM	LDSIL	LDMIL(Ours)	ROI	VBM	CLM	LDSIL	LDMIL(Ours)
AUC	0.8673	0.8762	0.8811	0.9574	0.9586	0.9178	0.9206	0.9537	0.9584	0.9716
ACC	0.7917	0.8050	0.8222	0.9056	0.9109	0.8696	0.8841	0.8986	0.9130	0.9275
SEN	0.7862	0.7735	0.7736	0.8742	0.8805	0.9130	0.9130	0.9783	0.9565	0.9348
SPE	0.7960	0.8300	0.8607	0.9303	0.9350	0.7826	0.8261	0.7391	0.8261	0.9130
F-Score	0.7692	0.7784	0.7935	0.8910	0.8974	0.9032	0.9130	0.9278	0.9362	0.9451
MCC	0.5800	0.6044	0.6383	0.8082	0.8191	0.7037	0.7391	0.7702	0.8018	0.8391