XAI-ML-AMD: Explainable Machine Learning for Android Malware Detection

This study presents a method for detecting malware in Android systems, addressing complex cyberattacks and advanced concealment techniques. It employs a range of machine learning approaches, including Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Logistic Regression, DecisionTree (DT), and Random Forest, combined with ensemble strategies such as stacking and voting to identify the most effective tool. To improve the model's clarity and reliability, the study employs Shapley Additive Explanations (SHAP) analysis. It also uses Local Interpretable Model-agnostic Explanations (LIME) to shed light on the logic driving specific detection outcomes, elucidating the process behind them. Through the examination of SHAP values and the creation of heatmaps, the research enhances the recognition of key or interrelated factors. Based on tests, the stacking ensemble method has the highest accuracy of 99.89%. This method combines strong detection with interpretable explanations, providing a reliable and practical means to counter real-world security challenges.