Smart Terminal Malware Detection Technology Based on Abnormal Network Behavior

Abstract

In recent years, the use of smart mobile terminals has increased rapidly, especially mobile phones on the Android platform, which account for almost half of the mobile phone market. Many attackers regard smart terminals as the target of their attacks. They can easily collect user privacy information, thus posing a huge threat to users. In this regard, researching a malware detection technology for smart terminals is of great significance to the safe use of mobile phones in my country. The purpose of this article is to study the smart terminal malware detection technology based on abnormal network behavior. In this paper, the method of reducing feature data is used to preprocess the captured original network data to improve the applicability of the naive Bayes algorithm in the network behavior data processing environment. The preprocessing part first cleans up the network data and removes useless information; then uses methods such as establishing a static address library and field query to divide the data, normalize the complex data, and finally construct the feature vector. The main part of the anomaly recognition module in this paper is the naive Bayes classifier, and the preprocessed data is introduced into the naive Bayes classifier as a feature vector for classification detection. This article introduces the privacy data monitoring technology to track the data flow of the malware detected by the anomaly recognition module, and determine the leakage path of its privacy data, so as to clarify the magnitude of its harmfulness for further processing. Research shows that this article has tested the performance of the system, and the time-consuming items all use the stopwatch timer that comes with the mobile phone to assist in timing, which can be accurate to 0.01 second.