Research Article
Towards Tax Evasion Detection Using Improved Particle Swarm Optimization Algorithm
Table 4
The results of the models based on different criteria.
| | Models | Precision | Recall | F1 | Accuracy |
| | SVM | 88.67 | 89.16 | 88.91 | 90.58 | | KNN | 81.70 | 82.42 | 82.06 | 85.33 | | C5.0 | 84.11 | 85.61 | 84.85 | 88.66 | | NB | 85.33 | 86.67 | 85.99 | 88.89 | | MLP | 89.82 | 90.45 | 90.13 | 91.33 | | Adaboost | 88.00 | 88.32 | 88.16 | 89.65 | | IPSO-MLP | 93.25 | 93.78 | 93.51 | 93.68 | | IPSO-SVM | 92.64 | 92.75 | 92.69 | 92.24 | | LR | 91.80 | 82.34 | 86.81 | 67.00 |
|
|
The bold value indicates the highest accuracy.
|