Cross-site Scripting Attack Detection using Combination of Multi-Layer Perceptron and Naive Bayes Algorithms

Today, the World Wide Web is the most widely used, cheapest, and fastest communication medium on the planet. Because of its accessibility, millions of individuals use it for their daily tasks. A website vulnerability is a fault or misconfiguration in a website's that allows an attacker to seize control of the site and maybe the hosting server. According to studies published in ۲۰۲۱ by the Open Web Applications Security Project, XSS attacks are a substantial danger to webapplications and the XSS vulnerability ranking third among the top ten vulnerabilities of web applications. To detect XSS attacks, a combination of two multilayer perceptron algorithms and a Naive Bayesian method was utilized in this research. Based on data derived from URLs and JavaScript code, the proposed method employs a hybrid machine learning algorithm to categorize regular and malicious web pages. Regarding the results of conducted experiments on a realworld data set, a combination of Naive Bayesian and multilayer perceptron algorithms had higher accuracy and precision than the other similar algorithms.

کلیدواژه ها:

Software Security ، XSS attacks ، Multi-Layer Perceptron ، Naive Bayes Algorithm ، Accuracy