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

Journal of Big Data

Table 5 The performance of the SVM classifier with the multi-manifold approach with step of 5 percent

Dataset	Recall	Precision	G-means	F-measure	Accuracy
ecoli1	0.90 ± 0.170	0.71 ± 0.228	0.88 ± 0.141	0.83 ± 0.159	0.87 ± 0.131
ecoli2	0.90 ± 0.202	0.88 ± 0.192	0.92 ± 0.131	0.90 ± 0.142	0.96 ± 0.062
ecoli3	0.93 ± 0.160	0.76 ± 0.270	0.93 ± 0.091	0.86 ± 0.195	0.94 ± 0.072
ecoli4	0.90 ± 0.200	0.78 ± 0.307	0.93 ± 0.137	0.92 ± 0.170	0.96 ± 0.071
ecoli0147vs56	0.85 ± 0.240	0.75 ± 0.311	0.80 ± 0.303	0.89 ± 0.119	0.88 ± 0.266
ecoli034_5	0.85 ± 0.320	0.63 ± 0.394	0.75 ± 0.387	0.88 ± 0.151	0.86 ± 0.261
ecoli0147_2356	0.78 ± 0.279	0.64 ± 0.327	0.75 ± 0.307	0.78 ± 0.211	0.86 ± 0.264
glass0	0.80 ± 0.165	0.67 ± 0.137	0.77 ± 0.099	0.79 ± 0.116	0.78 ± 0.098
glass0123456	0.98 ± 0.060	0.85 ± 0.244	0.93 ± 0.148	0.92 ± 0.166	0.91 ± 0.171
kddcup-buffer_overflow_vs_back	1 ± 0	0.88 ± 0.256	0.99 ± 0.014	0.95 ± 0.161	0.99 ± 0.027
new-thyroid1	0.91 ± 0.204	0.83 ± 0.274	0.93 ± 0.151	0.89 ± 0.206	0.94 ± 0.098
page-blocks-1-3_vs_4	0.57 ± 0.372	0.69 ± 0.440	0.66 ± 0.352	0.79 ± 0.262	0.94 ± 0.089
Pima	0.71 ± 0.163	0.58 ± 0.130	0.65 ± 0.131	0.63 ± 0.086	0.68 ± 0.122
segment0	0.97 ± 0.039	0.98 ± 0.023	0.98 ± 0.020	0.98 ± 0.017	0.99 ± 0.006
shuttle_2_vs_5	1 ± 0	1 ± 0	1 ± 0	1 ± 0	1 ± 0
vehicle2–1	0.97 ± 0.035	0.94 ± 0.079	0.98 ± 0.026	0.97 ± 0.033	0.98 ± 0.028
vowel0	0.95 ± 0.113	0.84 ± 0.198	0.96 ± 0.063	0.90 ± 0.120	0.97 ± 0.031
Wisconsin	0.98 ± 0.040	0.88 ± 0.079	0.90 ± 0.085	0.95 ± 0.053	0.91 ± 0.065
Average	0.881	0.793	0.872	0.879	0.912