Stacking Ensemble Learning Based Detection of Cross-Site Scripting Attack

In accordance with Cybersecurity statistics, a total average of 2,200 cyber-attacks occur per day. Cross-site Scripting (XSS) is one of the most notable weaknesses in web security. It belongs to the injection category. For an attacker to gain access to the user's systems the user has to interact with malicious links. This usually occurs when the victim visits any link from a website that may contain XSS payload or when visiting a vulnerable website that may trick the victim into sharing cookies resulting in a compromise of the security in the web application. In this regard, we concentrated on creating an advanced technique for identifying cross-site scripting attacks using the Stacking Ensemble Learning method is proposed. The model is implemented with a combination of different classifiers, namely Random Forest, Gaussian Naïve Bayes, SVM, and KNN, where Logistic Regression is used as the meta-model. This model has achieved a remarkable 99% accuracy rate while maintaining a high level of precision and Recall.