Research Article
Detecting Fraudulent Bank Account Based on Convolutional Neural Network with Heterogeneous Data
Table 3
Results of the classification methods.
| Method | No. | Optimum factor | Precision(%) | Sensitivity (%) | F1-score(%) |
| LOF | 1 | 0.1 | 23.71 | 25.42 | 24.54 | IForest | 2 | 0.08 | 64.39 | 57.46 | 60.73 | One-SVM | 3 | 0.08 | 32.75 | 41.87 | 36.75 | RC | 4 | 0.08 | 61.05 | 54.53 | 57.61 |
| NTD-CNN | 5 | | 81.48 | 73.33 | 77.19 | TTD-CNN | 6 | | 78.10 | 71.33 | 74.56 | HDF-CNN | 7 | | 85.14 | 76.8 | 80.76 |
|
|