Investigating rarity in web attacks with ensemble learners

Journal of Big Data

Table 9 Results for all seven classifiers and Brute Force web attacks across eight sampling ratios; Sampling column reports negative to positive class ratio after applying random undersampling, or “None” for case when no undersampling is applied; abbreviations are the same as in Table 8

Classifier	Sampling	AUC	SD AUC	—	Classifier	Sampling	AUC	SD AUC
LGB	None	0.52347	0.04591	—	DT	None	0.6081	0.01541
LGB	999:1	0.56732	0.07109	—	DT	999:1	0.71138	0.02993
LGB	99:1	0.82661	0.02345	—	DT	99:1	0.77387	0.0369
LGB	95:5	0.86462	0.02088	—	DT	95:5	0.81454	0.01794
LGB	9:1	0.89189	0.02004	—	DT	9:1	0.81491	0.02548
LGB	3:1	0.92633	0.01321	—	DT	3:1	0.92268	0.02977
LGB	65:35	0.9345	0.01298	—	DT	65:35	0.92906	0.01979
LGB	1:1	0.94182	0.01042	—	DT	1:1	0.92573	0.0118
RF	None	0.60966	0.02465	—	NB	None	0.71786	0.02205
RF	999:1	0.7004	0.04502	—	NB	999:1	0.71812	0.021
RF	99:1	0.79814	0.03736	—	NB	99:1	0.72041	0.02533
RF	95:5	0.81556	0.01969	—	NB	95:5	0.71591	0.01397
RF	9:1	0.82477	0.02217	—	NB	9:1	0.71825	0.01794
RF	3:1	0.91671	0.02861	—	NB	3:1	0.72329	0.02742
RF	65:35	0.94079	0.00935	—	NB	65:35	0.72421	0.02528
RF	1:1	0.9416	0.00757	—	NB	1:1	0.72745	0.02637
XGB	None	0.59958	0.01578	—	LR	None	0.5	0.0
XGB	999:1	0.66607	0.02099	—	LR	999:1	0.5	0.0
XGB	99:1	0.81074	0.02	—	LR	99:1	0.60291	0.01992
XGB	95:5	0.81301	0.02315	—	LR	95:5	0.60286	0.01719
XGB	9:1	0.81341	0.01607	—	LR	9:1	0.60202	0.01278
XGB	3:1	0.91119	0.03282	—	LR	3:1	0.69736	0.02262
XGB	65:35	0.93861	0.0095	—	LR	65:35	0.71535	0.04137
XGB	1:1	0.93797	0.00782	—	LR	1:1	0.81256	0.01301
CB	None	0.60474	0.0237	—
CB	999:1	0.61818	0.02931	—
CB	99:1	0.80572	0.02413	—
CB	95:5	0.81255	0.0217	—
CB	9:1	0.81618	0.02156	—
CB	3:1	0.8831	0.03923	—
CB	65:35	0.92989	0.01351	—
CB	1:1	0.9311	0.00945	—