Credit Card Fraud Detection on BankSim (Expected Cost)
22Expected CostFN
Evaluation Results
| Method | Links | |
|---|---|---|
| FNK=1.0, Labeled training sample size (N_l)=259462025.08 | 22 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=1.0, Labeled training sample size (N_l)=259462025.08 | 22.2 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=1.0, Labeled training sample size (N_l)=129732025.08 | 23.3 | |
| FNK=1.0, Labeled training sample size (N_l)=129732025.08 | 26.6 | |
| SVMK=1.0, Labeled training sample size (N_l)=259462025.08 | 27.8 | |
| RFK=1.0, Labeled training sample size (N_l)=259462025.08 | 28.9 | |
| SVMK=1.0, Labeled training sample size (N_l)=129732025.08 | 31.7 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=1.0, Labeled training sample size (N_l)=51902025.08 | 32.2 | |
| RFK=1.0, Labeled training sample size (N_l)=129732025.08 | 34 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=1.0, Labeled training sample size (N_l)=38932025.08 | 38 | |
| FNK=1.0, Labeled training sample size (N_l)=51902025.08 | 38.1 | |
| FNK=1.0, Labeled training sample size (N_l)=38932025.08 | 40.1 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=1.0, Labeled training sample size (N_l)=25952025.08 | 43.5 | |
| SVMK=1.0, Labeled training sample size (N_l)=51902025.08 | 43.5 | |
| FNK=1.0, Labeled training sample size (N_l)=25952025.08 | 45.4 | |
| RFK=1.0, Labeled training sample size (N_l)=51902025.08 | 53 | |
| SVMK=1.0, Labeled training sample size (N_l)=38932025.08 | 57.2 | |
| RFK=1.0, Labeled training sample size (N_l)=38932025.08 | 57.6 | |
| RFK=1.0, Labeled training sample size (N_l)=25952025.08 | 59 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.5, Labeled training sample size (N_l)=259462025.08 | 66.5 | |
| FNK=0.5, Labeled training sample size (N_l)=259462025.08 | 68 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.5, Labeled training sample size (N_l)=129732025.08 | 68.4 | |
| FNK=0.5, Labeled training sample size (N_l)=129732025.08 | 72 | |
| SVMK=0.5, Labeled training sample size (N_l)=259462025.08 | 72.4 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.5, Labeled training sample size (N_l)=51902025.08 | 73.6 | |
| SVMK=1.0, Labeled training sample size (N_l)=25952025.08 | 74.7 | |
| SVMK=0.5, Labeled training sample size (N_l)=129732025.08 | 74.9 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.5, Labeled training sample size (N_l)=38932025.08 | 77.8 | |
| FNK=0.5, Labeled training sample size (N_l)=51902025.08 | 78.5 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.5, Labeled training sample size (N_l)=25952025.08 | 79.8 | |
| FNK=0.5, Labeled training sample size (N_l)=38932025.08 | 81.9 | |
| L-SSLK=1.0, Labeled training sample size (N_l)=259462025.08 | 88.1 | |
| SVMK=0.5, Labeled training sample size (N_l)=51902025.08 | 88.6 | |
| FNK=0.5, Labeled training sample size (N_l)=25952025.08 | 96 | |
| LK=1.0, Labeled training sample size (N_l)=259462025.08 | 102.2 | |
| RFK=0.5, Labeled training sample size (N_l)=259462025.08 | 103.5 | |
| SVMK=0.5, Labeled training sample size (N_l)=38932025.08 | 107.4 | |
| L-SSLK=1.0, Labeled training sample size (N_l)=129732025.08 | 108.8 | |
| RFK=0.5, Labeled training sample size (N_l)=129732025.08 | 110.4 | |
| LK=1.0, Labeled training sample size (N_l)=129732025.08 | 124.6 | |
| RFK=0.5, Labeled training sample size (N_l)=51902025.08 | 127.2 | |
| L-SSLK=0.5, Labeled training sample size (N_l)=259462025.08 | 127.5 | |
| SVMK=0.5, Labeled training sample size (N_l)=25952025.08 | 128.9 | |
| RFK=0.5, Labeled training sample size (N_l)=38932025.08 | 130.7 | |
| L-SSLK=1.0, Labeled training sample size (N_l)=51902025.08 | 137 | |
| LK=0.5, Labeled training sample size (N_l)=259462025.08 | 137.2 | |
| L-SSLK=1.0, Labeled training sample size (N_l)=38932025.08 | 141.8 | |
| L-SSLK=0.5, Labeled training sample size (N_l)=129732025.08 | 142.8 | |
| SVMK=0.2, Labeled training sample size (N_l)=259462025.08 | 146.6 | |
| RFK=0.5, Labeled training sample size (N_l)=25952025.08 | 146.6 | |
| KNNK=1.0, Labeled training sample size (N_l)=259462025.08 | 148.4 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.2, Labeled training sample size (N_l)=38932025.08 | 149.7 | |
| SVMK=0.2, Labeled training sample size (N_l)=129732025.08 | 149.8 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.2, Labeled training sample size (N_l)=259462025.08 | 149.9 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.2, Labeled training sample size (N_l)=129732025.08 | 150.9 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.2, Labeled training sample size (N_l)=51902025.08 | 151 | |
| FNK=0.2, Labeled training sample size (N_l)=259462025.08 | 151.3 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.2, Labeled training sample size (N_l)=25952025.08 | 152.1 | |
| FNK=0.2, Labeled training sample size (N_l)=51902025.08 | 152.9 | |
| LK=1.0, Labeled training sample size (N_l)=51902025.08 | 157.5 | |
| FNK=0.2, Labeled training sample size (N_l)=38932025.08 | 159.1 | |
| SVMK=0.2, Labeled training sample size (N_l)=51902025.08 | 159.3 | |
| LK=0.5, Labeled training sample size (N_l)=129732025.08 | 159.4 | |
| FNK=0.2, Labeled training sample size (N_l)=129732025.08 | 161.1 | |
| L-SSLK=0.5, Labeled training sample size (N_l)=51902025.08 | 162.9 | |
| L-SSLK=1.0, Labeled training sample size (N_l)=25952025.08 | 168.1 | |
| SVMK=0.2, Labeled training sample size (N_l)=38932025.08 | 168.5 | |
| LK=1.0, Labeled training sample size (N_l)=38932025.08 | 169.4 | |
| L-SSLK=0.5, Labeled training sample size (N_l)=38932025.08 | 170.2 | |
| L-SSLK=0.2, Labeled training sample size (N_l)=259462025.08 | 171.1 | |
| LK=0.2, Labeled training sample size (N_l)=259462025.08 | 176.1 | |
| SVMK=0.2, Labeled training sample size (N_l)=25952025.08 | 182.5 | |
| KNNK=1.0, Labeled training sample size (N_l)=129732025.08 | 184.3 | |
| FNK=0.2, Labeled training sample size (N_l)=25952025.08 | 184.7 | |
| L-SSLK=0.2, Labeled training sample size (N_l)=129732025.08 | 184.9 | |
| LK=1.0, Labeled training sample size (N_l)=25952025.08 | 189.9 | |
| LK=0.5, Labeled training sample size (N_l)=51902025.08 | 191.1 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.1, Labeled training sample size (N_l)=38932025.08 | 192.9 | |
| FNK=0.1, Labeled training sample size (N_l)=259462025.08 | 193.1 | |
| L-SSLK=0.5, Labeled training sample size (N_l)=25952025.08 | 193.7 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.1, Labeled training sample size (N_l)=25952025.08 | 194.1 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.1, Labeled training sample size (N_l)=259462025.08 | 194.1 | |
| LK=0.2, Labeled training sample size (N_l)=129732025.08 | 195.4 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.1, Labeled training sample size (N_l)=129732025.08 | 197.1 | |
| SVMK=0.1, Labeled training sample size (N_l)=51902025.08 | 197.2 | |
| Semi-Supervised Bayesian GANs with Log-SignaturesK=0.1, Labeled training sample size (N_l)=51902025.08 | 198.6 | |
| SVMK=0.1, Labeled training sample size (N_l)=129732025.08 | 200.7 | |
| LK=0.1, Labeled training sample size (N_l)=259462025.08 | 203.3 | |
| LK=0.5, Labeled training sample size (N_l)=38932025.08 | 205.2 | |
| KNNK=0.5, Labeled training sample size (N_l)=259462025.08 | 205.9 | |
| L-SSLK=0.1, Labeled training sample size (N_l)=259462025.08 | 206.3 | |
| SVMK=0.1, Labeled training sample size (N_l)=38932025.08 | 209.2 | |
| SVMK=0.1, Labeled training sample size (N_l)=259462025.08 | 209.6 | |
| L-SSLK=0.1, Labeled training sample size (N_l)=129732025.08 | 209.9 | |
| L-SSLK=0.2, Labeled training sample size (N_l)=51902025.08 | 210.9 | |
| LK=0.1, Labeled training sample size (N_l)=129732025.08 | 216.1 | |
| L-SSLK=0.2, Labeled training sample size (N_l)=38932025.08 | 216.9 | |
| SVMK=0.1, Labeled training sample size (N_l)=25952025.08 | 218 | |
| FNK=0.1, Labeled training sample size (N_l)=129732025.08 | 218.4 | |
| RFK=0.2, Labeled training sample size (N_l)=259462025.08 | 221.7 |