Vaddikasulu kasani, Mahalakshmi Potluri, Naga Thanusha Thota, Kowshik Gedela, Mahesh Banne

Abstract

Android malware is an increasing threat in the field of mobile security, as the open-source nature and widespread use of the Android operating system make it a prime target for attackers. With the rapid growth of new malware variants, traditional detection techniques are no longer sufficient. Hence, there is a need for more intelligent and adaptive solutions. This paper proposes an enhanced malware detection framework that uses an optimized ensemble learning approach to improve the accuracy and efficiency of malware classification. The framework combines multiple machine learning models including Least Squares Support Vector Machine (LS-SVM), Logistic Regression (LR), Naive Bayes (NB), Decision Tree (DT), and K-Nearest Neighbours (KNN). Each of these classifiers contributes uniquely to the detection process, and by combining them through an ensemble method, the overall performance is significantly improved. The ensemble model applies a majority voting technique to determine th