Android Malware Detection And Family Classification using Permission Features With Intrusion Awareness

The relentless evolution and increasing intricacy of malware represent a formidable challenge to cybersecurity professionals. The presence of malware within a system poses a grave threat, encompassing a spectrum of potential calamities, from the theft of personal information to the sabotage of devices and networks. Therefore, it becomes imperative to devise solutions for the prompt and adaptive detection and classification of malware. This paper delves into the multifarious methods and techniques employed in the domain of Android malware detection and family categorization, with particular emphasis on identifying the permissions that could potentially lead to system intrusion. It also issues a warning message pertaining to the hazardous permissions that could potentially disrupt the system. By empowering users with this crucial information, the proposed system enable them to make informed decisions and take proactive measures to safeguard their systems. In a bid to ascertain the efficacy of our methods, rigorous testing of the dataset using a diverse range of machine learning models was conducted. The results were remarkable, with random forest for the identification of malware, adaboost for classification of family and isolation forest for identifying the permissions that can cause intrusion in the system.