DefinePK

Abstract

This paper presents a novel framework for detecting malware in Android applications using advanced machine learning techniques. Our approach combines static and dynamic analysis with deep learning algorithms to identify malicious patterns in Android applications. We propose a hybrid feature extraction method that captures both code- based and behavioral attributes, followed by a multi-layer classification model that achieves high detection accuracy. Experiments conducted on a comprehensive dataset of benign and malicious applications demonstrate the effectiveness of our approach, achieving 97.8% accuracy, 96.5% precision, and 98.2% recall. The proposed framework outperforms traditional signature-based methods and several existing machine learning approaches, showing promise for real-time malware detection in resource-constrained mobile environments