Research Article
A Deep Learning Method for Android Application Classification Using Semantic Features
Table 7
The results of the investigations.
| System | Dataset | Algorithm | ACC/F1 | Benign applications | Malware |
| Ours | Google play store: 18453 | VirusShare + drebin: 18549 | CNN | 99.3538%/99.3534% | [33] | Google play store: 3000 | Android malware genome project + Virusshare: 8000 | DBN | NA/95.05% | [14] | Google play store: 19747 | VirusShare: 13075 malgenome project: 1209 | DNN | 98%/NA | [34] | 360 security company and wandoujia app store:10948 | Drebin + VirusShare: 8652 | CNN | 96.54%/95.89% | [35] | AndroZoo: 5215 | VirusShare + android malware genome project: 5442 | DBN | 98.71%/NA | [36] | Google play store: 8000 | VirusShare: 8000 | CNN | 95.8%/NA |
|
|