Table 18 Detecting unknown malware families with the help of the PermDroid framework proposed in this study.
Combination of Android malware families to trained the model | Detection rate when trained PermDroid with DNN | Detection rat when trained PermDroi with NDTF |
---|---|---|
{A1, up to..... , A10} | 66% | 71% |
{A1, A3, up to..... , A11} | 70% | 69% |
: | : | : |
: | : | : |
: | : | : |
{A2, up to ..... , A11} | 59% | 55% |
: | : | : |
: | : | : |
: | : | : |
: | : | : |
: | : | : |
: | : | : |
{A7, up to ..... , A51} | 98.4% | 98% |
: | : | : |
: | : | : |
: | : | : |
: | : | : |
: | : | : |
: | : | : |