Effective and Explainable Detection of Android Malware Based on Machine Learning Algorithms

Abstract

The across the board reception of android devices and their ability to get to critical private and secret data have brought about these devices being focused by malware engineers. Existing android malware analysis techniques categorized into static and dynamic analysis. In this paper, we introduce two machine learning supported methodologies for static analysis of android malware. The First approach based on statically analysis, content is found through probability statistics which reduces the uncertainty of information. Feature extraction were proposed based on the analysis of existing dataset. Our both approaches were used to high-dimension data into low-dimensional data so as to reduce the dimension and the uncertainty of the extracted features. In training phase the complexity was reduced 16.7% of the original time and detect the unknown malware families were improved.