Research And Implementation Of Malicious Advertisement Detection System Based On Web

At present,online advertisements has become one of the indispensable economic assistance for the development of the Internet,but it has also become the first choice of computer hijacking technology by lawbreakers,which has brought huge economic losses and negative effects to the society.Due to the high dynamic variability and diversity of online advertising,as well as the development of evasive detection methods,there are some defects in the detection schemes based on advertising blacklist at present.Therefore,it is urgent to find a scheme that can effectively detect malicious advertisements.According to the formation principle of online advertisement,this thesis puts forward a new scheme to detect malicious advertisement,then designs and implements a detection system of malicious advertisement.Specific research results are as follows:1)aiming at the formation principle of online advertisements and the attack characteristics of malicious advertisements,this thesis proposes a malicious advertisement detection scheme based on the domain path of advertisement URL redirect chain combined with machine learning,which overcomes the problem that the attacker can change the lexical features to avoid detection.The detection scheme constructs the domain path of advertisement URL redirect chain and extracts some URL features.Finally,the malicious advertising detection model is tested on the test set.Through four comparative experiments and result analysis,it is proved that compared with current detection tools,the detection scheme proposed in this thesis has reached a relatively high level in ACC,FN and FP.2)In accordance with the detection scheme above and requirements analysis,this thesis designs and implements a malicious advertisement detection system which can be accessed by users through browser.The detection system includes five modules:advertisement extraction module,data preprocessing module,malicious advertisement detection module,web display module and task scheduling module.The advertisement extraction module using crawler technology to capture advertisement links,the data preprocessing module and malicious advertisement detection module using Python to develop function,the web display module is based on Angular2 framework and boostrap framework,the task scheduling module is developed based on SpringBoot framework,distributed file system FastDFS,quartz timer cluster and so on.The engineering code of the whole system is managed by git and gradle.Finally,by setting up several test cases,it verifies that the system has achieved the requirement analysis proposed by the system in terms of test results and performance efficiency.