Table 3.

Classification performance of different methods on ADNI-1

(a) AD versus NC
Method	ACC	SEN	SPE	AUC	P-value
Baseline-SVM	89.35±8.83	90.39	88.27	0.94	<0.001
LASSO-SVM	87.57±9.12	89.02	86.15	0.95	<0.001
t-test-SVM	86.75±10.33	83.92	89.42	0.93	<0.001
Baseline-MK-SVM	94.53±6.55	94.90	94.04	0.96	<0.001
LASSO-MK-SVM	93.74±7.81	95.00	91.60	0.97	<0.001
t-test-MK-SVM	93.45±7.35	94.90	91.92	0.96	<0.001
Group Lasso-MK-SVM	94.53±6.80	94.90	94.04	0.96	<0.001
Euclid-MK-SVM	95.08±6.77	97.25	92.88	0.97	0.004
Hypergraph-MK-SVM	94.77±6.39	97.25	92.31	0.97	<0.001
MFCC-MK-SVM	97.60±5.03	98.43	96.73	0.98	--
(b) MCI versus NC
Method	ACC	SEN	SPE	AUC	P-value
Baseline-SVM	70.75±10.04	79.80	53.46	0.76	<0.001
LASSO-SVM	72.46±11.05	83.03	52.31	0.78	<0.001
t-test-SVM	72.79±9.53	85.96	47.69	0.77	<0.001
Baseline-MK-SVM	80.09±8.24	87.47	65.96	0.79	<0.001
LASSO-MK-SVM	81.89±8.89	90.24	62.27	0.79	0.022
t-test-MK-SVM	81.71±9.43	91.82	62.31	0.79	0.019
Group Lasso-MK-SVM	79.76±6.91	95.76	49.23	0.77	<0.001
Euclid-MK-SVM	81.48±8.48	89.49	66.15	0.80	0.007
Hypergraph-MK-SVM	81.20±6.55	94.14	56.54	0.75	<0.001
MFCC-MK-SVM	84.47±6.83	94.04	66.15	0.81	--
(c) MCI-C versus MCI-NC
Method	ACC	SEN	SPE	AUC	P-value
Baseline-SVM	53.95±15.12	44.65	61.07	0.59	<0.001
LASSO-SVM	54.57±14.87	45.12	61.79	0.60	<0.001
t-test-SVM	50.76±13.74	34.42	63.39	0.57	<0.001
Baseline-MK-SVM	69.17±12.77	57.44	78.04	0.66	<0.001
LASSO-MK-SVM	71.88±13.36	61.97	76.00	0.66	<0.001
t-test-MK-SVM	63.05±12.60	50.70	72.32	0.59	<0.001
Group Lasso-MK-SVM	70.86±11.37	62.33	77.14	0.65	<0.001
Euclid-MK-SVM	72.00±12.97	69.77	73.57	0.70	<0.001
Hypergraph-MK-SVM	73.64±11.19	66.28	79.11	0.74	0.008
MFCC-MK-SVM	77.76±10.59	67.44	85.54	0.76	--