Intelligent Mobile Malware Detection

8System Call Pattern-Based Detection

DOI: 10.1201/9781003121510-8

In the previous chapters, we employed machine learning classifiers for malware detection. Machine learning mechanisms usually require good feature representation for accurate classification of the data points. It is a challenge to identify and represent the correct features of the data points. Further machine learning based detection mechanisms can be computationally challenging and may not be suitable for on device real time deployment. Hence, there have been many attempts to detect malware based on various signatures.

According to Artenstein et al. [77], malware applications can utilize the vulnerabilities in the Android IPC to perform various kinds of attacks such as sending ...