ANDROID MALWARE DETECTION WITH RUNTIME FEATURES

Abstract

The increasing sophistication and increasing sophistication of Android applications makes malware detection increasingly difficult. Obfuscation techniques used by malware developers reduce the effectiveness of static analysis because these techniques can successfully mask flaws. In response to the limitations of static analysis, recent research has focused on dynamic analysis techniques that are more resistant to obfuscation. The CCCS-CIC-AndMal-2020 file was published by the Canadian Cyber Security Institute and contains various signatures from Android malware. These features are divided into six categories: memory, network, battery, engine, processing, and API. Personal impact of malware classification category. This study aims to fill this gap by examining the impact of each type of optimization on Android malware detection. We use a variety of filters and wrappers to identify and report on the most important categories and highlight important features within those categories. This comprehensive review not only improves our understanding of the role of dynamic features in malware detection, but also provides insights into improving process deployment. Selection, classification and obfuscation flexibility in malware detection, CCCS-CIC-AndMal-2020 dataset.