A Comparative Analysis of Android Malware Detection with and without Feature Selection Techniques using Machine Learning

Abstract

Android is an open-source operating system mainly built for smart devices to make them easy to use and user-friendly. Thus, it has immensely engulfed other operating systems in mobile devices, which have become not only a major stakeholder in the market but have also become attractive targets for cyber criminals to lure many Androids malware with the intention of stealing or destroying the user's information without the user knowing. Many traditional signature-based anti-malware efforts have been made to combat malicious apps, but these efforts have been insufficient due to the lack of ability to detect unknown malware. This insufficient effort by traditional signature-based has led to the intervention of researchers to embark upon combating unknown malware using machine learning techniques. This study looks into many existing research papers on malware detection using machine learning in order to determine the significance of feature selection techniques. The comparative analysis examines the importance of feature selection and unselected feature techniques