Back to Auxiliary material page
SVM-light vs. ELR (for only UCI binary-class datasets)
	svm-light		svm-light		svm-light		svm-light		svm-light		svm-light		svm-light		svm-light
	best value		t1d2		c0.5t1d2		c0.05t1d2		c0.005t1d2		c0.0005t1d2		c0.00005t1d2		c0.000005t1d2
australian	77.1	2.88	60.15	6.21	64.93	14.1	70.29	9.11	66.38	11.76	66.38	13.56	57.97	15.05	77.1	2.88
breast	96.62	1.23	96.62	1.23	93.53	2.23	93.97	1.21	94.41	2.36	96.47	1.42	96.62	1.23	96.32	1.56
chess	98.97	0	95.97	0	98.97	0	97.65	0	94.47	0	91.09	0	54.78	0	53.19	0
cleve	80.34	3.08	65.42	6.06	73.9	3.08	72.54	4.39	72.88	2.68	79.32	4.7	78.98	3.08	80.34	3.08
corral	100	0	99.2	1.79	100	0	96.8	5.22	84.8	7.16	56.8	10.35	56.8	10.35	56.8	10.35
crx	70.31	6.43	57.54	5.53	70.31	6.43	70.15	8.34	68.61	10.17	68.77	10.51	64	15.34	60.77	15.94
diabetes	76.34	3.5	75.69	2.9	68.89	5.48	69.28	5.77	71.89	6.96	75.16	4.05	76.34	3.5	75.86	3.97
flare	82.91	3.13	82.91	3.13	82.63	3.02	82.06	3.81	82.82	3.4	82.91	3.13	82.91	3.13	82.91	3.13
german	68.7	5.75	68.7	5.75	66.1	2.41	66.2	1.75	65.6	2.99	66.3	0.97	66.5	1.12	66.4	2.16
glass2	79.37	8.45	53.75	11.35	73.75	5.68	79.37	8.45	74.38	11.35	71.88	12.3	66.25	10.69	62.5	11.89
heart	83.33	3.21	64.82	7.75	76.67	2.11	76.67	2.81	77.04	3.61	79.63	4.54	80.81	5.02	83.33	3.21
hepatitis	86.25	5.23	83.75	7.13	86.25	5.23	86.25	5.23	86.25	5.23	86.25	5.23	86.25	5.23	83.75	3.42
mofn-3-7-10	100	0	100	0	100	0	100	0	77.34	0	77.34	0	77.34	0	77.34	0
pima	75.95	2.03	74.9	3.08	70.2	4.2	70.59	4.03	70.33	5.12	75.68	2.97	75.95	2.03	74.77	2.67
vote	95.17	1.5	95.17	1.5	92.87	1.71	93.1	1.15	95.4	1.41	93.79	3.31	61.38	8.32	61.38	8.32
average	84.757		78.306		81.267		81.661		78.84		77.85		72.192		72.851
NB+ELR>SVM-light(best average)				p<	0.02264
TAN+ELR>SVM-light(best average)				p<	0.00362
GBN+ELR>SVM-light(best average)				p<	0.0078
best average: c=0.05 + poly 2 (t=1 d=2)
							svm-light
	NB+ELR		TAN+ELR		GBN+ELR		best value (cXt1d2)
australian	84.93	1.06	84.93	1.03	86.81	1.11	77.1	2.88
breast	96.32	0.66	96.32	0.7	95.74	0.43	96.62	1.23
chess	95.4	0.64	97.19	0.51	90.06	0.92	98.97	0
cleve	81.36	2.46	81.36	1.78	82.03	1.83	80.34	3.08
corral	86.4	3.25	100	0	100	0	100	0
crx	86.46	1.85	86.15	1.7	85.69	1.3	70.31	6.43
diabetes	75.16	1.39	73.33	1.97	76.34	1.3	76.34	3.5
flare	82.82	1.35	83.1	1.29	82.63	1.28	82.91	3.13
german	74.6	0.58	73.5	0.84	73.7	0.68	68.7	5.75
glass2	81.88	3.62	80	3.9	78.75	3.34	79.37	8.45
heart	78.52	3.44	78.15	3.86	78.89	4.17	83.33	3.21
hepatitis	86.25	5.38	85	5.08	90	4.24	86.25	5.23
mofn-3-7-10	100	0	100	0	100	0	100	0
pima	75.16	2.48	74.38	2.58	74.25	2.53	75.95	2.03
vote	95.86	0.78	95.4	0.63	95.86	0.78	95.17	1.5
average	85.408		85.921		86.05		84.757
NB+ELR>SVM-light (best value)				p<	0.35082
TAN+ELR>SVM-light (best value)				p<	0.19453
GBN+ELR>SVM-light (best value)				p<	0.1964
best value: the c value for best accuracy + poly 2
-c float    - C: trade-off between training error
and margin (default [avg. x*x]^-1)
-t int      - type of kernel function:
0: linear (default)
1: polynomial (s a*b+c)^d
2: radial basis function exp(-gamma ||a-b||^2)
-d int      - parameter d in polynomial kernel