Android Malware Detection Method Based on Multi-Features Fusion

Abstract

Machine learning methods for Android malware detection usually encounter the problem of single feature extraction, high feature dimension and low detection accuracy. This paper proposes an android malware detection method based on multi-features fusion through statistic analysis on the authority and category information of android software. The proposed method effectively combines the authority features with other category features and utilizes information gain to extract features with high class discrimination and low dimension. Compared with other methods, the proposed method can improve the detection accuracy of machine learning methods in malware detection through multi-features fusion.