Research On Security Analysis Techniques Of Android Applications

Android intelligent terminal plays an important role in our daily life.But in recent years,it also faces many security challenges and poses a great threat to users.How to detect the security of Android application accurately and efficiently is a hot spot of current research.This paper first introduces the security mechanism and threats of Android,then summarizes the research status of three aspects in the field of Android application security detection,including static analysis technology,dynamic analysis technology and machine learning analysis technology.On the basis of this,a new static and dynamic integrated detection method for Android applications in this paper is proposed.It includes two parts,which are the detection of heavily packaged malware based on the matching of identifier information and the multidimensional Fuzzing vulnerability mining based on Taint Analysis.Finally,the effectiveness of the detection method is proved by the experimental simulation.In addition,this paper also proposes an improved SVM-KNN classification method for solving the existing problems of SVM-KNN algorithm in parameter optimization and feature weight calculation.From the results of static and dynamic analysis,we collect all kinds of characteristic data,such as risk permissions,risk mechanisms,sensitive calls API and malicious behavior,to do equalization selection and dimension reduction.The interpolation search method is used to improve the mesh optimization method in support vector machine(SVM).The eigenvalues are calculated by the eigenvalue method and the K nearest neighbor method(KNN)is combined with Mahalanobis distance.In the experimental simulation stage,the comparison results show that this method has higher classification accuracy and shorter search time.Finally,this paper designs and realizes an Android application security detecting system(AASDS)with three layers detection model based on the proposed detection method.It divides into two parts of client and server.It can detect most malicious applications and vulnerabilities that harm the users of Android system,and gives some targeted suggestions.