Research And Implementation Of Android Malicious Code Detection Technology Based On Behavior Analysis

With the increase in the number of phone users as well as the development of the Android system,till the end of December 2015,the number of Android applications has reaches to fifteen million.With the increasing malicious applications,how to detect potential malicious applications is the key and difficult to promote the mobile internet.In this paper,android malicious code detection technology based on behavior analysis using machine learning methods for analyzing application behaviors detecting of potentially malicious applications,but with the mobile application increasing,selecting the appropriate detection model has become one of the key problems we propose a method of malicious code detection based on behavior analysis.Because the Extreme Learning Method(ELM)has the features such as fast learning skill,strong generalization ability and high accurate rate,which is widely used in data mining,computer vision,pattern recognition.So this paper will show you how to use ELM in the behavior of malicious code detection process.The main work of this paper is as follows:1)Doing research and analysis of the mobile application behavior detection technology and ELM.2)Propose the potential malicious application detection method based on ELM.The method includes:Behavior acquisition,Feature extraction,Feature optimization and Training of the ELM.This method makes up the disadvantages of traditional machine learning methods,such as high cost and long training time.3)Develop a malicious code detection system.It extracts behavior characteristics and optimizes the feature from the potential malicious applications.Then the existing models ELM will classify it.4)With the normal application and malicious application as an example,the malicious code detection system was tested.The experimental results show that the method proposed in this paper has good detection accuracy.