A Core Set Based Large Vector-Angular Region and Margin Approach for Novelty Detection

Table 3

Training time on different datasets.

Dataset	ν-SVM (s)	SVDD (s)	MAMC (s)	Core set based LARM (s)
Australian	0.0739 ± 0.0086	0.5975 ± 0.0993	0.6325 ± 0.0122	0.8819 ± 0.4613
B. authentication	0.1627 ± 0.0699	0.3441 ± 0.0199	1.3943 ± 0.0131	1.1103 ± 0.1658
B. Cancer	0.0975 ± 0.0063	0.2462 ± 0.0152	0.7309 ± 0.0126	2.2633 ± 0.2803
Cod-rna	388.7901 ± 24.6651	NULL	NULL	7.4578 ± 0.1622
Covtype	NULL	NULL	NULL	64.9519 ± 7.3727
Diabetic	0.2096 ± 0.0063	1.1188 ± 0.0819	1.9243 ± 0.0119	5.7415 ± 1.0748
Fourclass	0.0354 ± 0.0030	0.1271 ± 0.0151	0.2706 ± 0.0117	0.5652 ± 0.1175
Glass	0.0255 ± 0.0038	0.0587 ± 0.0038	0.1459 ± 0.0088	0.2079 ± 0.0220
Heart	0.0229 ± 0.0016	0.0858 ± 0.0082	0.1148 ± 0.0070	0.8305 ± 0.2195
H. valley	0.1457 ± 0.0119	1.0385 ± 0.0278	1.4355 ± 0.0164	0.8506 ± 0.2926
Ionosphere	0.0634 ± 0.0097	0.1999 ± 0.0134	0.4018 ± 0.0087	1.9758 ± 0.3947
L. disorders	0.0399 ± 0.0058	0.1861 ± 0.0202	0.1809 ± 0.0095	1.6280 ± 0.4039
MC	19.5417 ± 0.3331	49.0111 ± 10.8097	472.2837 ± 1.3679	17.1713 ± 5.3223
SDD	13.8246 ± 1.0316	145.7426 ± 18.0463	218.5634 ± 0.6386	8.3123 ± 0.5485
S. segmentation	NULL	NULL	NULL	7.2536 ± 0.3645
Sonar	0.0463 ± 0.0050	0.1406 ± 0.0061	0.1896 ± 0.0054	2.3453 ± 0.4481
Shuttle	226.0020 ± 3.1471	268.0641 ± 145.7989	NULL	7.3061 ± 0.2336
Svmguide1	0.4634 ± 0.0171	1.3081 ± 0.3742	8.9181 ± 0.0539	1.0625 ± 0.0781
Wilt	2.0750 ± 0.0774	8.6527 ± 0.7402	52.8582 ± 0.7763	3.6379 ± 0.3202
Wine	0.0263 ± 0.0049	0.0983 ± 0.0274	0.1785 ± 0.0111	0.2936 ± 0.0494