Android Software Ad Plugin Security Research And Detection System Implementation

With the rapid development of mobile internet industry,the android mobile smart terminal become more popular,the android mobile applications have constantly varied,mobile application content has covered all aspects of people's lives.The needs of user's mobile applications,makes more developers to join the ranks of application development,also that embedded advertising by means of mobile applications and reap the benefits via advertising makes the business model become more and more popular.With the mobile application advertising industries have been expanding,mobile applications advertising is facing more and more security risks.There are some security issues to mobile application advertising,such as traffic consumption,malicious deductions,privacy theft and other problems.So it is necessary to analyze the security of the advertising libraries in application,and detect advertising libraries in application.First of all,this paper summarizes the current research status of the advertising libraries.Then,this paper analyzes the security problem of advertising libraries and the detection method of advertising libraries.In terms of security analysis of the advertising libraries,the main analysis focuses on the use of permission and function call,and many researches only focus on one aspect,a few of researchers analyze security problems of all aspect.This paper analyze the security vulnerabilities and the risk in advertising libraries,the collection of user information,the use of permission,and other bad behaviors.Finally,according to the analysis results,aim at the detection of advertising libraries this paper propose a thread which based on the number of words.Through the research of the existing detection methods of advertising libraries,which mainly includes the method based on the white list and the method of machine learning based on semantic analysis.After analyzed the advantages and disadvantages of that two methods,this paper propose a method of detect advertising libraries that based on decompile sequence of numbers.Firstly,this method decompiles the application,then counts word for each file that from decompile,and records the path information of files,according to the form of advertising package name and path information to split the module.Then we can learn the word number sequence obtained of each module.On the basis of this word number sequence and through the machine learning methods to achieve the detection of advertising libraries.This word number sequence avoid the situation of getting less semantic features,but also solve whitelist method inadequate.By comparing the experimental,the result shows that the feature has very good performance in terms of detect advertising libraries,with a higher accuracy rate.At last,according to this method,this paper designs and implements advertising libraries detection system and this system achieves high accurate detection of advertising libraries.