An Empirical Study on the Performance of Cost-Sensitive Boosting Algorithms with Different Levels of Class Imbalance

Table 7

Index of performance using 𝐺-mean for data sets with a low-level imbalance.

	Win	Loss	Index	Win	Loss	Index	Win	Loss	Index	Win	Loss	Index	Win	Loss	Index
	0.1			0.2			0.3			0.4			0.5
AdaC1	22	0	22	18	0	18	13	0	13	8	0	8	6	0	6
AdaC2	6	8	−2	3	7	−4	4	7	−3	3	4	−1	9	2	7
AdaC3	0	14	−14	0	11	−11	2	6	−4	10	2	8	15	1	14
AdaCost	6	1	5	3	0	3	2	2	0	1	6	−5	0	12	−12
CSB1	7	0	7	3	2	1	0	5	−5	0	6	−6	0	8	−8
CSB2	0	18	−18	0	7	−7	0	1	−1	0	4	−4	0	7	−7
	0.6			0.7			0.8			0.9			1
AdaC1	12	0	12	12	0	12	10	0	10	11	0	11	15	0	15
AdaC2	17	1	16	19	1	18	19	0	19	20	0	20	15	0	15
AdaC3	18	1	17	21	2	19	19	0	19	17	1	16	13	0	13
AdaCost	0	23	−23	0	24	−24	0	24	−24	0	24	−24	0	24	−24
CSB1	0	13	−13	0	15	−15	0	16	−16	0	17	−17	1	17	−16
CSB2	0	9	−9	0	10	−10	0	8	−8	0	6	−6	0	3	−3