Research Article
A Low Computational Cost Method for Mobile Malware Detection Using Transfer Learning and Familial Classification Using Topic Modelling
Table 7
Comparison of training time and computational power overhead.
| Data set | Model | Mean time to detect (MTTD) in seconds | Computational power requirements |
| Hybrid | CNN | 90000 | Processor: Intel Xeon, GPU: Nvidia RTX 3080, RAM: 128 GB |
| Malgenome | Transfer learning | 61200 | Processor: Intel Core i9 10900K, GPU: Nvidia RTX 3060, RAM: 96 GB |
| AAGM | Transfer learning | 55944 | Processor: Intel Xeon, GPU: Nvidia RTX 3060, RAM: 96 GB |
|
|