Breast Cancer Detection in the IOT Health Environment Using Modified Recursive Feature Selection

Table 5

Classification results of SVM (kernel = RBF)-based predictive model on different features subsets created by REF FA algorithm.

Model	Classification performances evaluation metrics
	Parameters (C, )	Number of features	Acc (%)	Spe (%)	Precision (%)	Sen/recall (%)	F1-score	MCC	Classification error (%)	Execution time (s)
SVM (RBF)	(1, 0.0001)	1	64	100	99	3	6	50	36	0.005
		2	64	100	98	5	10	50	36	0.008
		3	84	100	98	56	72	78	16	0.006
		4	85	100	98	63	77	81	15	0.006
		5	85	100	98	62	77	80	15	0.006
		6	86	100	99	62	77	80	14	0.007
		7	86	100	99	62	77	81	14	0.006
		8	86	100	99	62	77	82	14	0.006
		9	87	100	98	63	77	82	13	0.007
		10	87	100	98	63	77	82	13	0.007
		11	87	100	98	63	77	82	13	0.007
		12	91	99	99	77	87	88	9	0.005
		13	90	98	98	76	86	87	10	0.005
		14	91	99	99	77	87	88	9	0.004
		15	92	99	99	81	89	91	8	0.008
		16	90	100	100	76	87	87	10	0.017
		17	92	98	98	86	92	91	8	0.004
		18	98	99	99	96	98	97	2	0.004
		19	97	98	98	96	98	97	3	0.003
		20	97	99	99	96	98	97	3	0.004
		21	97	99	99	95	97	97	3	0.005
		22	97	99	99	95	97	97	3	0.004
		23	95	99	99	89	94	94	5	0.008
		24	94	99	99	89	94	94	6	0.006
		25	94	99	99	89	94	94	6	0.015
		26	94	99	99	89	94	94	6	0.009
		27	94	99	99	89	94	93	6	0.016
		28	95	98	98	88	93	94	5	0.018
		29	94	99	99	89	94	94	6	0.017
		30	95	99	99	90	95	94	5	0.019