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

Table 1

Datasets used in experiments.

Dataset	#pos	#neg
Australian	383	307	192	11	14
Banknote authentication (B. authentication)	762	610	381	19	4
Breast Cancer (B. Cancer)	458	241	229	22	9
Cod-rna	39690	19845	19845	99	8
Covtype	283301	211840	141651	141	54
Diabetic	611	540	306	15	19
Fourclass	307	555	154	7	2
Glass	144	70	72	7	11
Heart	120	150	60	6	13
Hill valley with noise (H. valley)	299	307	150	15	100
Ionosphere	225	126	113	5	34
Liver disorders (L. disorders)	200	145	100	10	6
Magic gamma telescope (MC)	12332	6688	6166	12	10
Sensorless drive diagnosis (SDD)	5319	53190	2660	26	48
Skin segmentation (S. segmentation)	194198	50859	97099	97	3
Sonar	111	97	56	5	60
Shuttle	34108	9392	17054	17	9
Svmguide1	2000	2000	1000	10	4
Wilt	4265	74	2133	10	5
Wine	119	59	84	8	13