Research And Implementation Of Malicious Code Identification

At present, Internet users spend a lot of time to visit the web, but a large number of websites contain malicious code more or less. When the user browses the websites, the malicious code will be added to the user's computer system, so that his computer system unknowingly infected with the malicious code or computer viruses.Some knowledge of how to identify the website of malicious code are introduced in this thesis.It used the artificial intelligence advanced method-machine learning,nalysed elements of work by many kinds of malicious code in homepages,and privde some measures to prevente them.On the Web,more and more malicious code appear.In order to identify the malicious code,we use the machine learning in a wide range of examples and mechanical combination.The keywords of some common malicious code are used the machine learning as example.It estalish the explain of learning and judge foot-language of the Web,and then separate all the code of Javascript of the website,and compare it and the keywords,so that remove the unsafe code or suspicious code.Machine learning of the system through the implementation unit compare javascript program segment with repository.If the code is the same with or similar to the feature keywords, remove the segment from the original code after extracting feature keywords from similar code through study unit and storing these new examples in the repository to enhance the ability of learning system self-study.While the code which does not match the repository, it is considered as safe code and is reserved. So it can detecte and filtere for homepage which contains the malicious code into the browser, to reduces its damage.