The Research Of Malicious Detection Based On Multi-dimensional Technology In Android

The detection of malicious application has been the most interesting part of Android System Protection.With the development of code obfuscation technology,it has been more difficult to detect malicious applications.At the same time,the situation is more serious with many users ignore the request of Permissions.Currently,the detection of malicious application for Android is focused on behavior,but most studies only consider a single data source for feature extraction.Therefore,we need an effective method to inspect and analyze the malicious of applications to help users found the problems exist in system.For the system of Android,the raise of a fast and accurate detection for malicious applications programs need consider how to address the following issues:How to use and improve the previous proposed methods,making design more suitable for mobile computing?How to avoid the impact of confusion,make it easy to extract,analyze and study malicious features?How to measure the Malicious features?For these questions,the main work of this paper as follows:1.This paper has a detailed study of the protection of Android system,main components,Kirin protection strategy and Apex framework.We analyze the principle and trends of malware applications,and study the main malicious detection of android.2.By analyzing all the Permissions of the Android and the correlation between them,we obtained the combinations of permission that may disclose user information and attack system.Assigning a weight for each combination according to a known set of benign and malicious applications could be used to indicate the degree of danger.3.According to the study of protection strategy and framework,we modify the original code of installer of Android system,and we use the theory of combination to enhance the analysis of permission.The results of test are displayed in form of digital and level,allowing users to be aware of potential threats to the application and make reasonable judgments.4.Repackage has become the main way to produce malicious application.We propose a feature extraction rely on representative files to detect malicious applications effectively.This method is not only resistant to the impact of confusion,but also can be used with the detection of relevant-permission.This kind of multi-feature extraction could be further improve the accuracy of detection.