The Trust Model For The Behavior Of The Android App

With the rapid progress of the traditional Internet technology,the mobile communication technology and the entire IT industry,the mobile internet technology develops rapidly,each industry is affected by the information age.Especially in recent years,the network technology extends to the direction of broadband,so the mobile e-commerce,mobile instant messaging,mobile social,mobile payment and other new business based on the mobile Internet begin to integrate into our daily live.Mobile Internet authentication of end-user identity,the protection of user-sensitive information,the safety supervision of mobile Internet and other security issues become more prominent and a focus of attention.Android system is currently the most used of a good open source mobile platform,its application scope and attention has been increasing.Therefore,its security issues have become the focus of the industry and users.This thesis carried out the research from the perspective of detecting the Android platform malware.First of all,this thesis introduced the system structure,operating mechanism and the existing security measures of the Android platform.Secondly,this thesis analyzed the advantages and disadvantages of the method and mechanism about detection.Then,on the basis of extracting the Android malicious program features,the thesis established a set of effective model about app malware behavior trust.Finally,the thesis used this model to detect the behavior of app Android.The results were as follows:1.Determine the objectives about behavior assessment.Using Androguard and other anti assembly tools to analysis the behavior characteristics and the code structure of the software,and using these two aspects as the inspection index;2.A feature extraction algorithm is proposed.VSM vector model is used to extract and quantify the characteristics of malicious software,and the formal feature vectors can be effectively used;3.Build an app behavior trust model.Using the Analytic Hierarchy Process(APH),combined with the above feature vector extraction and formal methods,a set of effective app behavior trust model is established.Experimental results show that the established trust model can effectively identify the malicious software and improve the detection ability of the malicious software on Android platform.