SmartShield: Harnessing Code Extraction and ML for Next-Gen Malware Detection

The rapid proliferation of Android applications has led to an increased prevalence of malware, posing significant security risks to users and devices. This research presents a novel Hybrid Model for Android malware detection, integrating multiple machine learning algorithms to enhance accuracy and robustness. The methodology involves the systematic collection of a comprehensive dataset, extraction of critical features using the Code Extraction Principle (CEP), and the evaluation of various models including SVM, Random Forest, Gradient Boosting, XGBoost, and others. The proposed Hybrid Model significantly outperforms these models, achieving an accuracy of 97.81%, with precision and recall rates of 96.72% and 95.89%, respectively. This model’s superior performance is attributed to the effective combination of different algorithms, optimizing feature extraction, and the robust handling of both benign and malicious applications. The model’s robustness against evasion techniques further underscores its potential as a reliable tool in Android malware detection. The results indicate that the proposed Hybrid Model is highly effective in minimizing both false positives and false negatives, offering a promising solution for enhancing mobile security. The findings of this research contribute to the development of more secure and efficient malware detection systems, addressing the evolving challenges in mobile cybersecurity.