An Explainable Software Engineering Metric-Based Ensemble Framework for Android Malware Detection
Year: 2026
Type: other
Abstract: This study presents an explainable machine learning framework for Android malware detection based on Software Engineering (SE) metrics. The proposed approach extracts 89 SE features from Android applications and applies explainability-driven feature selection using LIME and SHAP to identify the most influential characteristics associated with malicious behavior. Three machine learning models—Random Forest (RF), Multilayer Perceptron (MLP), and Logistic Regression (LR)—are combined using a soft-voting ensemble strategy to improve detection performance and robustness. The framework is evaluated on an Android malware dataset containing 4,711 applications, including 3,048 benign and 1,663 malicious samples. Experimental results demonstrate that the proposed ensemble approach achieves high classification performance while maintaining interpretability through explainable feature analysis. The study highlights the effectiveness of integrating Software Engineering metrics, ensemble learning, and Explainable Artificial Intelligence (XAI) techniques for reliable and transparent Android malware detection. The proposed framework can support cybersecurity applications by providing both accurate malware classification and insights into the software characteristics contributing to detection decisions.
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7202294366