Research Article
MFEMDroid: A Novel Malware Detection Framework Using Combined Multitype Features and Ensemble Modeling
Table 4
Results based on different networks or classifiers.
| Network | Accuracy (%) | Precision (%) | Recall (%) | F1-score (%) | Mcc (%) | Training time (min) |
| KNN | 92.31 | 96.28 | 86.96 | 91.37 | 84.84 | — | DT | 93.05 | 93.61 | 91.40 | 92.49 | 86.05 | — | SVM | 92.20 | 93.27 | 89.84 | 91.51 | 84.37 | — | ResNet | 93.99 | 94.74 | 92.28 | 93.50 | 87.93 | 0.46 | SENet | 94.57 | 95.02 | 93.28 | 94.14 | 89.09 | 0.44 | SEResNet | 95.20 | 96.65 | 93.16 | 94.87 | 90.42 | 0.45 | MFEMDroid | 95.38 | 95.78 | 94.47 | 95.12 | 90.74 | 3.79 |
|
|
Bold values signify the best results.
|