Table 10 Comparison of ECS-SDE with five deep ensemble models.

From: Example dependent cost sensitive learning based selective deep ensemble model for customer credit scoring

Datasets

Metrics

ECS-SDE

LSTM-GRU-ANN

LSTM-GRU-MLP

CNN- BLSTM

BiLSTM-CNN

BiLSTM-Trans-CNN

GMSC

Save

0.45448(1)

(± 0.01887)

0.45255 (3)

(± 0.02754)

0.44879 (5)

(± 0.02872)

0.45048 (4)

(± 0.01435)

0.33459 (6) (± 0.03692)

0.45638 (2) (± 0.01839)

AUC-PR

0.24449(1)

(± 0.01451)

0.15741 (4)

(± 0.01324)

0.15519 (5)

(± 0.03435)

0.17215 (3)

(± 0.02324)

0.18976 (2) (± 0.00619)

0.13061 (6)

(± 0.01836)

AUC-ROC

0.80102(1)

(± 0.03872)

0.75952 (2)

(± 0.03567)

0.75256 (3)

(± 0.02446)

0.74638 (4)

(± 0.01223)

0.67733 (6)

(± 0.03597)

0.74343 (5)

(± 0.01299)

BS+

0.21223 (1)

(± 0.03133)

0.27363 (3)

(± 0.04436)

0.29424 (4)

(± 0.04136)

0.36674 (5)

(± 0.02326)

0.54003 (6) (± 0.04525)

0.22046 (2) (± 0.04880)

BS–

0.15019 (1)

(± 0.04304)

0.20732 (5)

(± 0.04643)

0.20064 (4)

(± 0.03216)

0.15951 (3)

(± 0.02346)

0.15531 (2) (± 0.02334)

0.31267 (6)

(± 0.04900)

PAKDD

Save

0.30556(1)

(± 0.03316)

0.11141 (5)

(± 0.01242)

0.15683 (4)

(± 0.02436)

0.19230 (2)

(± 0.03437)

0.14678 (6)

(± 0.03504)

0.16268 (3)

(± 0.04154)

AUC-PR

0.25155(1)

(± 0.02`197)

0.24275 (4)

(± 0.03448)

0.25110 (2)

(± 0.009423)

0.25000 (3)

(± 0.04154)

0.24029 (5) (± 0.01476)

0.23869 (6)

(± 0.01394)

AUC-ROC

0.60951(1)

(± 0.04021)

0.58504 (4)

(± 0.02547)

0.60083 (2)

(± 0.03216)

0.59762 (3)

(± 0.01346)

0.58203 (5) (± 0.00775)

0.57823 (6) (± 0.01637)

BS+

0.18299(1)

(± 0.03758)

0.49906 (6)

(± 0.04141)

0.47649 (4)

(± 0.03221)

0.44828 (2)

(± 0.03456)

0.46803 (3)

(± 0.04274)

0.49530 (5)

(± 0.01012)

BS–

0.38199 (3)

(± 0.03635)

0.39007 (5)

(± 0.02350)

0.33086 (2)

(± 0.03222)

0.32827 (1)

(± 0.02336)

0.39791 (6)

(± 0.04960)

0.38824 (4)

(± 0.00877)

DCCC

Save

0.33307(1)

(± 0.01751)

0.30498 (4)

(± 0.01772)

0.30975 (3)

(± 0.03316)

0.32182 (2)

(± 0.02408)

0.18496 (6)

(± 0.03701)

0.20921 (5)

(± 0.04914)

AUC-PR

0.41299(1)

(± 0.01704)

0.36568 (3)

(± 0.022082)

0.35958 (4)

(± 0.03879)

0.38707 (2)

(± 0.03567)

0.34615 (5) (± 0.01274)

0.33523 (6)

(± 0.04805)

AUC-ROC

0.72550(1)

(± 0.03672)

0.68748 (4)

(± 0.03678)

0.69414 (3)

(± 0.03213)

0.70537 (2)

(± 0.01193)

0.65016 (6)

(± 0.03955)

0.65645 (5) (± 0.04450)

BS+

0.22276(1)

(± 0.02338)

0.45140 (3)

(± 0.02662)

0.38660 (2)

(± 0.01239)

0.52629 (5)

(± 0.04272)

0.57824 (6)

(± 0.04396)

0.49509 (4)

(± 0.01365)

BS–

0.12378 (2)

(± 0.04118)

0.35958 (6)

(± 0.02126)

0.17363 (4)

(± 0.03309)

0.10163 (1)

(± 0.01723)

0.12944 (3)

(± 0.04771)

0.19201 (5) (± 0.04931)

IEEE

Save

0.51258(1)

(± 0.03616)

0.50320 (3)

(± 0.03021)

0.50170 (4)

(± 0.02301)

0.51007 (2)

(± 0.01331)

0.41981 (5) (± 0.04353)

0.41855 (6)

(± 0.02332)

AUC-PR

0.50040(1)

(± 0.01391)

0.23584 (2)

(± 0.04931)

0.22872 (3)

(± 0.02323)

0.18457 (4)

(± 0.03351)

0.14851 (5)

(± 0.01191)

0.13404 (6)

(± 0.04439)

AUC-ROC

0.86714(1)

(± 0.03963)

0.83397 (3)

(± 0.02911)

0.84520 (2)

(± 0.01903)

0.82097 (4)

(± 0.03421)

0.77659 (5)

(± 0.02285)

0.59283 (6)

(± 0.03782)

BS+

0.36915 (1)

(± 0.04371)

0.36961 (2)

(± 0.03211)

0.37029 (3)

(± 0.04951)

0.37155 (4)

(± 0.02361)

0.38498 (6) (± 0.04353)

0.37763 (5) (± 0.03430)

BS–

0.01314 (1)

(± 0.01094)

0.08652 (4)

(± 0.03343)

0.06074 (2)

(± 0.01208)

0.06999 (3)

(± 0.01991)

0.10719 (5)

(± 0.01191)

0.11256 (6)

(± 0.00584)

Average ranking

1.15

3.75

3.25

2.95

4.95

4.95

Back to article page