The Research And Design Of Dynamic Detection Of Malicious Apllication Based On Softwares' Behavior

There are various malicious applications (app) in mobile platform, especially for Android devices. As the new classes and features have been appeared on malwares,it is difficult to get a clear malware's description.There are limited conditions of triggering malwares to performance malicious behaviors as well as a record of malwares' behaviors can be none-malicious,in a result, it's hard to construct a model to detect malwares with malicious application directly and also difficult for those kinds of model to detect the newest malwares effectively. It's important to design a kind of malware detection method which can be highly efficient and accurate.This article begin with the introduction of basic knowledge of security mechanism of Android system, then havea mention of its existing problems, and state an' analysisof the differences between variety malicious software, and also have a conclude of the method and the algorithm used in malware detection. On the basis of that,this paper proposes a dynamical detection method of malicious software based on RRNS. Thi's paper also presents a method to improve the performance of detection algorithm by using sample generation algorithm. Finally, the related experiments and evaluation are carried out, and the existing problems are summarized, and the future development of malicious software detection methods is discussed.