Research On XSS Attack Detection Technology Based On Machine Learning

With the advent of the Web2.0 era,the client's interactions with the server are more and more,and the interactions are more convenient and there are a lot of Web attacks at the same time.Cross-Site Scripting(XSS)is one of the most common and most harmful Web attack,it is a serious threat to the security of clients,servers,and databases.Many testing tools for XSS attacks on the market are almost static,and are difficult to defend against unknown and deformed XSS attacks.In order to improve the detection effect of the unknown and the diversified XSS attack of tools,the Hidden Markov Model(HMM)and Support Vector Machine(SVM)two kinds of Machine learning algorithms are used to detect XSS attack in this paper,this paper makes the following research.First of all,the research status and background of XSS attack detection are briefly described,in order to improve the effect of the traditional tools for detecting the diversified XSS,machine learning algorithm the importance of detecting XSS by machine learning algorithm is proposed.The method and diversification XSS attack are summarized by the analysis of XSS attack samples,and a theoretical basis is provided for the feature extraction of two machine learning models.Secondly,the model principle of HMM,learning problems and evaluation problems involved in this paper are illustrated.Then the word set model is used to simulate the XSS attacks sample feature extraction,the feature is used to generate the HMM model as observed sequence,the forward-backward algorithm is used to calculate the threshold value and the output value,and the testing of the model is complete.The HMM detector does better than Chrome XSS-Filter in terms of accuracy,rate of falseand rate of missing,and it performs better than Chrome XSS Filter on diversified XSS.Finally,the SVM algorithm is used for XSS detection.The principle of SVM model is illustrated,and how to use kernel function to solve the nonlinear SVM problem is analyzed.Then the feature of XSS samples is analyzed and extracted,and the appropriate kernel function is selected to train nonlinear SVM,The SVM detector also does better than Chrome XSS-Filter in terms of accuracy,rate of false and rate of missing.