Optimized Convolutional Neural Networks based malware detection

Abstract

Malware is a malicious code which is developed to harm a computer or network. The number of
malwares is growing so fast and this amount of growth makes the computer security researchers
invent new methods to protect computers and networks. It is a very serious problem and many
efforts are devoted to malware detection in today’s cybersecurity world. To detect malware, most
antivirus scanners use a combination of signature matching and detection based on exploration.
The obvious problem with this is that only known and lesser-known specimens can be identified
based on the artifacts extracted from malware analysis. This is no longer the case due to the rapid
growth of malware in recent years. In this thesis, we have proposed an optimized malware
detection method for the Android platform using an optimized approach based on optimized
neural networks, which are both a combination of permissions associated with risk and
vulnerable API calls. And uses them as features of the CNN algorithm. The results show that the
accuracy of the proposed method is 93.75%, which is 2.92% better than the basic article method.