AdaBoost-based multiple SVM-RFE for classification of mammograms in DDSM

Table 6 Comparison of methods by maximum Az value using 8 features (Calcification)

	T	MGH	WU	WFUSM	SHH
SVM		0.91182	0.98765	0.94690	0.96595
SVM-RFE		0.91196	1.00000	0.95100	0.96595
ENSEMBLE		0.53915	0.69512	0.56583	0.91392
JOIN (1)		0.67508	0.71655	0.83947	0.93422
JOIN (2)		0.57971	0.72941	0.76157	0.88733
JOIN (3)		0.57971	0.72941	0.62686	0.73542
JOIN (4)		0.54571	0.69512	0.62686	0.72464
JOIN (5)		0.54571	0.69512	0.54077	0.66210
MSVM-RFE (bootstrap)	5	0.91182	0.98765	0.95326	0.97868
	10	0.91182	0.98765	0.95168	0.96595
	15	0.91182	0.98765	0.94690	0.96757
	20	0.91182	0.98765	0.94690	0.97348
MSVM-RFE (boost)	5	0.91182	0.98765	0.94690	0.96595
	10	0.91182	0.99259	0.94690	0.96595
	15	0.91182	0.99429	0.94690	0.96595
	20	0.91182	0.98765	0.94690	0.96595

ISSN: 1472-6947