Android Malware Classification using XGBoost based on Images Patterns
Huajun Chen, Ranjin Du, Zhen Liu, Huan Xu · 2018 IEEE 4th Information Technology and Mechatronics Engineering Conference (ITOEC) · 2018
With the rapid development of mobile Internet, mobile terminals gradually become the national basic information equipment. As the most popular mobile operating system, Android has many extr, emely serious security issues. In our wok, we extracted the dex file from the Android malware and visualized it as the image to extract features. Depend on the massive Android apps, we used XGBoost as the classification models to expand experiments. Compared with KNN, ET and GBDT, the XGBoost adopted the best classification effect, reaching the 99.14% accuracy and the 99.10% recall. The experimental results testing on a malware sets containing 10 families demonstrate that XGBoost has the excellent performance for Android malware classification.