Android System Malicious Code Detection And Defense Technology Research

The rapid development of the current mobile network,especially the rapid rise of the current 3G and 4G networks,mobile devices are increasingly becoming the work of life essential tools。The popularity of mobile devices has also brought a series of problems,especially security issues are the most difficult problems.The safety of mobile devices is increasingly being tested.Android system with the advantages of its open source that favored by various mobile device manufacturers,from the first Android system intelligent machine released,just two years to account for the global 48% market share.As of the second quarter of 2016,Android mobile phone market share has reached 86.2%,a record high.The ratio is more higher than IOS(12.9%)and Windows phone(0.9%).Faced with such data,Android system security can not be ignored.Based on the research of scholars at home and abroad,this paper studies the malicious code detection and defense technology of Android system in combination with the current popular technology.Firstly,this paper introduces the research status of the malware detection and defense technology of Android system,analyzes the current situation and the security situation of the malicious code in Android system,and the research results in this field.Analysis of the difficulties and challenges in this field.A detailed analysis of the Android system architecture,the Android system architecture and its security mechanism to do an in-depth analysis.This paper introduces the popular machine learning disciplines,analyzes the definition of learning,feature selection criteria,classification algorithm,small sample statistics theory and so on.Finally,according to the characteristics of Android system malicious code,this paper proposes a technology of based on classes-SVM malware detection and defense technology of Android system.This paper gives a theoretical explanation and gives the realization scheme of malicious code detection and defense technology based on classes-SVM in Android system.Based on the advantages of the current machine learning disciplines,with the characteristics of malicious code in Android system,and the SVM machine learning algorithm is selected.According to the same set of features in APP in the same class,if there is an exception to the set of features in a class of APP,we can predict the existence of malicious code in the APP.According to this principle,design scheme,the first artificial intervention classification of Android system APP,decompile the data,select the feature two set of permissions and API set to create the model for training,selection of SVM classification algorithm to analyzed,finally the relevant assessment of data model,finally come to the conclusion that the proposed scheme outperforms other SVM machine study of detection technology,proves the feasibility of this technology.At present,there is no authoritative malicious APP sample library,the sample of this paper from the major APP market and well-known sites as test data,and test the relevant training model.Test results show that,based on the classes-SVM Android system malicious code detection and defense technology is better.The program meets the expected results.