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

Feizi, Tayyebe; Moattar, Mohammad Hossein; Tabatabaee, Hamid

doi:10.1186/s40537-023-00832-2

Journal of Big Data

Table 12 The average precision of different under-sampling methods

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

Dataset	Original	RUS	NCL [28]	OSS [27]	CNN [34]	ENN [35]	CBU [31]	PUMD [13]	Proposed
ecoli1	0.78 (4)	0.83 (2)	0.83 (2)	0.78 (4)	0.77 (6)	0.77 (6)	0.77 (6)	0.89 (1)	0.72 ± 0.212 (9)
ecoli2	0.80 (7)	0.86 (3)	0.82 (6)	0.76 (8)	0.76 (8)	0.86 (3)	0.86 (3)	0.90 (1)	0.88 ± 0.192 (2)
ecoli3	0.66 (6)	0.77 (2)	0.55 (9)	0.58 (8)	0.70 (5)	0.65 (7)	0.76 (3)	0.88 (1)	0.76 ± 0.270 (3)
ecoli4	0.73 (7)	0.79 (4)	0.76 (6)	0.71 (9)	0.82 (3)	0.73 (7)	0.88 (2)	0.90 (1)	0.78 ± 0.307 (5)
ecoli0147vs56	0.84 (2)	0.78 (8)	0.84 (2)	0.83 (5)	0.80 (6)	0.84 (2)	0.80 (6)	0.90 (1)	0.75 ± 0.311 (9)
ecoli034_5	0.83 (3)	0.80 (7)	0.82 (5)	0.83 (3)	0.86 (2)	0.82 (5)	0.80 (7)	0.90 (1)	0.79 ± 0.335 (9)
ecoli0147_2356	0.73 (7)	0.77 (2)	0.76 (4)	0.77 (2)	0.71 (8)	0.74 (6)	0.75 (5)	0.90 (1)	0.66 ± 0.253 (9)
glass0	0.51 (8)	0.67 (2)	0.60 (5)	0.60 (5)	0.51 (8)	0.59 (7)	0.67 (2)	0.83 (1)	0.67 ± 0.137 (2)
glass0123456	0.79 (8)	0.60 (9)	0.82 (5)	0.81 (6)	0.83 (3)	0.80 (7)	0.83 (3)	0.89 (1)	0.86 ± 0.227 (2)
kddcup-buffer_overflow_vs_back	0.15 (9)	0.89 (3)	0.82 (5)	0.80 (7)	0.80 (7)	0.81 (6)	0.89 (3)	0.9 (2)	1 ± 0 (1)
new-thyroid1	0.78 (8)	0.78 (8)	0.82 (5)	0.81 (6)	0.86 (3)	0.86 (3)	0.81 (6)	0.81 (2)	0.93 ± 0.155 (1)
page-blocks-1-3_vs_4	0.88 (4)	0.83 (8)	0.88 (4)	0.88 (4)	0.91 (1)	0.82 (9)	0.86 (5)	0.90 (2)	0.90 ± 0.200 (2)
Pima	0.62 (8)	0.67 (5)	0.68 (3)	0.72 (2)	0.63 (7)	0.66 (6)	0.68 (3)	0.75 (1)	0.61 ± 0.148 (8)
segment0	0.87 (8)	0.87 (8)	0.88 (5)	0.88 (5)	0.89 (3)	0.88 (5)	0.89 (3)	0.90 (1)	0.98 ± 0.023 (2)
shuttle_2_vs_5	0.91 (3)	0.91 (3)	0.91 (3)	0.91 (3)	0.91 (3)	0.91 (3)	0.89 (9)	0.9 (2)	1 ± 0 (1)
vehicle2–1	0.74 (5)	0.80 (3)	0.73 (6)	0.72 (8)	0.75 (4)	0.73 (6)	0.83 (2)	0.71 (9)	0.94 ± 0.079 (1)
vowel0	0.82 (6)	0.85 (4)	0.81 (8)	0.86 (3)	0.82 (6)	0.81 (8)	0.89 (2)	0.9 (1)	0.85 ± 0.185 (4)
wisconsin	0.86 (5)	0.84 (8)	0.87 (3)	0.85 (6)	0.87 (3)	0.84 (8)	0.85 (6)	0.9 (2)	0.93 ± 0.083 (1)
Average Ratings	6.06 (9)	5.00 (6)	4.83 (4)	5.28 (7)	4.89 (5)	5.83 (8)	4.39 (3)	1.72 (1)	3 (2)
Average Precision	0.74 (9)	0.80 (4)	0.79 (6)	0.78 (8)	0.79 (6)	0.78 (5)	0.82 (3)	0.87 (1)	0.83 (2)

Back to article page