SECURING ANDROID DEVICES: A ROBUST APPROACH TO AUTOMATEDMALWARE DETECTION THROUGH ENSEMBLE LEARNING
Keywords:
Android security, malware detection, machine learning, ensemble learning, cybersecurity, zero-day attacks, behavioral analysis, feature engineering, transfer learning, real-time threat detection, false positive reduction, support vector machines (SVMs), decision trees, neural networks, static and dynamic analysis, API call analysis, Permission based detection, artificial intelligence (AI), deep learning, federated learning, data privacy, Internet of Things (IoT) security, mobile edge computingAbstract
The growing complexity of Android malware presents serious security threats, necessitating sophisticated detection methods beyond the capabilities of conventional antivirus tools.In order to improve accuracy and efficiency, this paper proposes a solid method for automated malware detection using ensemble learning, which combines several machine learning classifiers. To categorize Android apps as either benign or malicious, we specifically use Random Forest, Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM). Through the utilization of Soft Voting in a Voting Classifier, our model combines all classifiers' predictions to enhance detection accuracy and reduce false positives. Experimental results indicate that our ensemble learning solution effectively improves malware detection performance with significant improvement, and it offers an effective and scalable security solution for Android devices.
