Advancing Android Security: Leveraging Stacking Ensemble and Bioinspired Feature Selection for Efficient Malware Detection
V. Jyothsna, Peddesugari Mokshitha, Shaik Khulud, L. Reddy, Nare Jagannath Reddy, Bhasha Pydala
Abstract
Nowadays, the use of smartphones and related applications is increasing rapidly due to their “easy to use” features, the functionality of many applications, and the constant development of software and hardware of smart devices. Research in this field shows that 4.5 billion people will have mobile phones by 2024. The most used smartphone operating system among these devices is Android. Android accounts for 75.5% of the market. Because Android is so widely used, it is more vulnerable to viruses and malware, making it an ideal target for hackers. Android anti-malware technology, which can quickly detect and classify different malware to create a rapid response plan, has become popular in recent years. The usefulness of combining deep learning and machine learning to offer automated and self-learning services has been shown by numerous applications; nevertheless, the shortage of malware samples has been identified as a barrier to the development of deep learning-based solutions. To solve this problem, this work introduces new machine-learning techniques that can classify malware into different groups and identify malware attacks. The proposed method used the Drebin dataset divided into hazardous and non-hazardous substances.