A binary classification model that detects whether the given JS program is obfuscated or not.
Dataset is taken from https://www.kaggle.com/datasets/fanbyprinciple/obfuscated-javascript-dataset
| Obfuscated | Not Obfuscated |
|---|---|
| 1477 | 1857 |
The features were trained using Weka and the obtained results as follow.
| ML Algorithm | Accuracy (45 features) | Accuracy (47 features) |
|---|---|---|
| Naive Bayes | 94.47% | 96.20% |
| Logistic Regression | 98.47% | 98.76% |
| Support Vector Machine | 97.65% | 99.13% |
| Decision Tree | 99.05% | 99.13% |
| Ada Boost | 99.01% | 99.17% |
| Lazy k-star | 98.97% | 99.22% |
| Boosted Decision Tree | 99.17% | 99.30% |
| Random Forest | 99.20% | 99.55% |