Table 10. Results of the proposed SVO & SVOX methods on the Adult9a dataset with different values of k.

k	Reduced (ratio)		Training Speedup		Testing Speedup		Accuracy		Number of SVs		Jaccard similarity
Whole dataset	1.00		1.00		1.00		0.850		11580		1.00
	SVO	SVOX	SVO	SVOX	SVO	SVOX	SVO	SVOX	SVO	SVOX	SVO	SVOX
1	0.37	0.22	4.54	25.38	1.36	5.10	0.701	0.844	8678	2845	0.58	0.24
2	0.38	0.23	4.67	23.57	1.36	5.10	0.710	0.843	8768	2839	0.59	0.24
3	0.38	0.23	4.52	23.23	1.44	5.20	0.718	0.844	8858	2845	0.60	0.24
4	0.39	0.23	4.55	24.47	1.41	5.25	0.725	0.844	8925	2859	0.60	0.24
5	0.39	0.24	4.40	23.09	1.39	5.18	0.731	0.843	9003	2857	0.61	0.24
10	0.40	0.25	4.20	20.30	1.32	4.85	0.756	0.845	9259	2923	0.63	0.25
15	0.41	0.26	4.00	19.95	1.33	5.03	0.770	0.845	9445	2953	0.65	0.25
20	0.42	0.27	3.76	20.02	1.29	5.09	0.781	0.844	9562	2979	0.66	0.25
30	0.44	0.28	3.56	17.05	1.24	4.89	0.794	0.844	9775	3006	0.68	0.25
40	0.45	0.29	3.37	16.50	1.20	4.92	0.801	0.844	9907	3031	0.69	0.26
50	0.46	0.31	3.13	16.04	1.15	4.97	0.808	0.844	10047	3049	0.71	0.26

Each row compares performance metrics of the methods for a different value (1 to 50) of the parameter k of the SVO & SVOX methods. The first row represents the metrics measured after training on the whole dataset. “Reduced (ratio)” is the ratio of the size of the resulting reduced dataset to the size of the whole dataset.