A multi-manifold learning based instance weighting and under-sampling for imbalanced data classification problems

Journal of Big Data

Table 16 Average F-measure of different under-sampling methods on kddcup datasets

Dataset	Manifold model	RBUS [48]	CRIUS [1]	NBUS [13]	NN_HIDC [45]	Proposed
kddcup-buffer_overflow_vs_back	multi-manifold	0.66	0.83	0.43	1	1 ± 0
kddcup-rootkit-imap_vs_back	single-manifold	0.83	0.86	0.78	1	1 ± 0
kddcup-guess_passwd_vs_satan	multi-manifold	0.00	0.00	0.00	0.99	0.99 ± 0.027
kddcup-land_vs_portsweep	single-manifold	0.96	0.23	0.23	0.23	1 ± 0
kddcup-land_vs_satan	single-manifold	0.97	0.97	0.97	1	1 ± 0
Average F-measure		0.685(3)	0.578(4)	0.482(5)	0.844(2)	0.998(1)