Table 4.

Classification performance of different methods on ADNI-2

(a) AD versus NC
Method	ACC	SEN	SPE	AUC	P-value
Baseline-SVM	91.13±5.04	92.37	89.50	0.95	<0.001
LASSO-SVM	85.90±5.51	89.34	81.38	0.92	<0.001
t-test-SVM	79.60±6.93	84.31	73.38	0.86	<0.001
Baseline-MK-SVM	91.72±4.15	93.36	89.56	0.94	0.006
LASSO-MK-SVM	86.82±4.57	89.57	82.66	0.90	<0.001
t-test-MK-SVM	90.06±4.35	92.75	86.50	0.93	<0.001
Group Lasso-MK-SVM	89.92±4.42	93.65	85.00	0.93	<0.001
Euclid-MK-SVM	91.72±4.15	93.36	89.56	0.94	0.006
Hypergraph-MK-SVM	91.19±4.12	94.17	87.25	0.94	<0.001
MFCC-MK-SVM	93.72±3.38	95.17	91.81	0.95	--
(b) LMCI versus NC
Method	ACC	SEN	SPE	AUC	P-value
Baseline-SVM	69.23±7.25	74.46	63.37	0.74	<0.001
LASSO-SVM	66.61±6.60	71.66	60.96	0.71	<0.001
t-test-SVM	62.81±6.12	70.38	54.28	0.65	<0.001
Baseline-MK-SVM	74.35±5.99	81.42	66.42	0.77	<0.001
LASSO-MK-SVM	71.46±6.00	76.86	62.72	0.71	<0.001
t-test-MK-SVM	73.00±5.76	81.52	63.42	0.75	<0.001
Group Lasso-MK-SVM	74.35±6.15	81.42	66.42	0.77	<0.001
Euclid-MK-SVM	74.35±5.99	81.42	66.42	0.77	<0.001
Hypergraph-MK-SVM	75.32±5.79	85.07	64.39	0.75	<0.001
MFCC-MK-SVM	78.47±5.61	85.88	70.16	0.78	--
(c) EMCI versus LMCI
Method	ACC	SEN	SPE	AUC	P-value
Baseline-SVM	64.08±6.79	76.48	45.99	0.66	<0.001
LASSO-SVM	63.55±7.13	78.32	42.03	0.66	<0.001
t-test-SVM	63.32±5.35	87.33	28.29	0.64	<0.001
Baseline-MK-SVM	70.01±5.52	85.20	47.86	0.68	<0.001
LASSO-MK-SVM	68.43±4.83	88.92	37.31	0.66	<0.001
t-test-MK-SVM	69.10±5.25	85.05	45.83	0.66	<0.001
Group Lasso-MK-SVM	70.22±4.40	90.62	40.43	0.68	<0.001
Euclid-MK-SVM	70.01±5.52	85.20	47.86	0.68	<0.001
Hypergraph-MK-SVM	71.45±4.43	90.95	42.99	0.68	0.001
MFCC-MK-SVM	73.87±4.77	90.55	49.52	0.70	--