Research Article
What Is the Internet Water Army? A Practical Feature-Based Detection of Large-Scale Fake Reviews
Table 4
Classification result on JD dataset.
| ā | Accuracy | AUC | F1 | Precision | Recall |
| AdaBoost | 0.630 | 0.680 | 0.629 | 0.632 | 0.628 | Neural network | 0.636 | 0.703 | 0.628 | 0.658 | 0.583 | Gradient boosting | 0.666 | 0.720 | 0.662 | 0.669 | 0.646 | Logistic regression | 0.645 | 0.698 | 0.622 | 0.664 | 0.588 | Naive Bayes | 0.589 | 0.668 | 0.399 | 0.742 | 0.274 | Random forest | 0.651 | 0.730 | 0.691 | 0.665 | 0.659 | SVM | 0.606 | 0.679 | 0.471 | 0.722 | 0.363 | Ensemble | 0.630 | 0.704 | 0.571 | 0.698 | 0.455 |
|
|