Klasifikasi Malware Android Dengan Menggunakan Metode XGBoost Algoritma

Android, an operating system developed by Google, dominates the global market with a share of 71.8% by the end of 2023. While this success is driven by its open-source nature and the variety of applications on the Google Play Store, Android is also a major target for malware attacks, with 97% of all malware attacks in 2022 targeting Android devices. Malware also continues to improve over time, making it increasingly difficult to detect. Therefore, a reliable detection method is needed. In today's IT field, machine learning has shown quite efficient results in detecting malware. The author proposes the XGBoost Algorithm method as an approach to classifying Android malware. In this study, the Feature Selection technique was applied, namely Recursive Feature Elimination (RFE) and Multicollinearity Removal (MR), to reduce data dimensions and improve model performance. Testing was carried out by comparing the performance of the XGBoost model before and after the application of Feature Selection. The evaluation results using the classification report and confusion matrix show that the XGBoost model that applies Feature Selection successfully achieves Validation Accuracy of 98%, Detection Accuracy of 98%, Precision of 98%, Recall of 98%, and F1-Score of 98%. The model without Feature Selection only achieves a value of 97% on the same metrics.