Research Article
Effective and Efficient Android Malware Detection and Category Classification Using the Enhanced KronoDroid Dataset
Table 5
Results after hyperparameter tuning.
| | Task | Metric | RF (%) | DT (%) | KNN (%) | SVM (%) |
| | Malware detection | Accuracy | 98.03 | 96.60 | 97.16 | 96.39 | | Precision | 98.51 | 97.12 | 97.24 | 97.65 | | Recall | 97.73 | 96.45 | 97.44 | 95.54 | | F1-score | 98.12 | 96.78 | 97.34 | 96.58 |
| | Malware category classification | Accuracy | 87.56 | 83.09 | 84.92 | 84.78 | | Precision | 90.14 | 74.14 | 77.27 | 78.40 | | Recall | 74.20 | 70.87 | 73.67 | 71.34 | | F1-score | 81.39 | 72.47 | 75.20 | 74.70 |
|
|
Bold values are highest results/values of Random Forest classifier/detector.
|