Research On Malware Detection Technology For Android System

The open and free Android operating system gets favor of the majority of users.However,with the rapid growth of the market share,Android operating system gradually attracts the attention of hackers.The malware of Android system turns to be a serious threat to the rights and interests of users,research on malware detection technology for Android system has become a hot topic in the field of information security.On the basis of researching and analyzing the existing work,this paper conducts a research on malware detection technology for Android operating system,the main contribution of this paper is as follows.1.In order to improve the classification accuracy of Android application further,a malware integration detection scheme for Android system is proposed based on binary particle swarm,na?ve Bayesian,hierarchical na?ve Bayesian and integrated learning algorithm.The Android application sample database is classified according to the function at first,and then the feature extraction,processing and selection are applied to each application sample,and finally,the selected optimal privilege feature subset is used to train the classifier.The integrated detection scheme can improve the classification accuracy of classifier for malware in Android.2.Based on the Android malware integration detection scheme proposed in this paper,a feature selection algorithm for Android application based on binary particle swarm and na?ve Bayesian is designed.In order to improve the ability of Na?ve Bayesian algorithm to guide the binary particle swarm optimization algorithm,this paper deals the conditional probability in the traditional na?ve Bayesian algorithm,and then uses the packaging method to encapsulate the na?ve Bayesian algorithm been smooth and the binary particle swarm optimization algorithm,optimizes the permission feature set selection for the original Android application,and choose optimal privilege feature subset for the Android application.The experimental results show that the feature selection algorithm can choose the optimal privilege feature subset,which can effectively improve the classification accuracy of Android application classification algorithm.3.Based on the Android malware integration detection scheme proposed in this paper,an Android application classification algorithm based on hierarchical na?ve Bayesian and Bagging is designed in the classifier construction stage.The hierarchical na?ve Bayesian algorithm is used to train the optimal set of privilege features selected in the feature selection stage at first,and a number of hierarchical na?ve Bayesian classifiers are constructed,then the Bagging method is used to integrated learn from the hierarchical na?ve Bayesian classifier,and build strong classifier BHNB,which effectively guarantee the stability of Android application classifier,and improve the classification accuracy of Android application classification algorithm further.4.Based on the proposed Android malware integration detection scheme,feature selection algorithm and classification algorithm,this paper designs and implements the malware prototype detection system BMDroid.The experimental results show that the detection system can effectively detect malware,and compared with the Droid Mat?Androguard?DroidScope and Crowdroid,the detection accuracy of proposed detection system is higher.