Research Article
AndroClass: An Effective Method to Classify Android Applications by Applying Deep Neural Networks to Comprehensive Features
Table 8
Improvement in classification effectiveness (%) with manual dataset over real-world datasets.
| ā | Google | APKPure | ā | accuracy | precision | recall | F-score | accuracy | precision | recall | F-score |
| KNN | 88.1 | 196.4 | 240.4 | 220.6 | 125.6 | 119.5 | 148.3 | 135.1 | NB | 70.5 | 203.1 | 218.5 | 211.8 | 93.9 | 108.6 | 154.8 | 132.8 | SVM | 73.3 | 175.8 | 170.8 | 173.9 | 104.2 | 136.3 | 125.8 | 131.1 | DNN | 59.6 | 143.0 | 139.7 | 141.3 | 80.3 | 108.3 | 97.1 | 102.2 |
|
|