Table 9.

Performance and comparison results of the proposed predictor on the five NASA datasets (cost-sensitive case, four different cost ratios).

NASA datasets	Performance metrics	CR (cost ratio) =  costfalse_negative/costfalse_positive
		CR = 4.00		CR = 1.50		CR = 0.67		CR = 0.25
		ANN-ABC [5]	Proposed model	ANN-ABC [5]	Proposed model	ANN-ABC [5]	Proposed model	ANN-ABC [5]	Proposed model
CM1	pd (or) TPR	91	96	75	81	62	74	21	32
	pf (or) FPR	60	45	38	32	28	24	5	3
	Accuracy	46	61	64	72	71	79	88	93
JM1	pd (or) TPR	99	99	80	87	44	56	8	21
	pf (or) FPR	94	87	54	56	17	12	2	2
	Accuracy	24	35	53	76	76	88	81	97
KC1	pd (or) TPR	93	98	87	93	59	74	22	45
	pf (or) FPR	58	41	44	39	20	18	4	3
	Accuracy	50	76	61	84	77	90	85	97
KC2	pd (or) TPR	90	96	80	92	74	87	37	67
	pf (or) FPR	39	35	27	24	19	14	5	3
	Accuracy	67	78	74	86	80	94	84	97
PC1	pd (or) TPR	94	98	89	96	66	78	23	41
	pf (or) FPR	48	32	36	27	23	20	2	2
	Accuracy	55	62	66	79	76	87	93	95