Web Intrusion Classification System using Machine Learning Approaches
Mansi Bhatnagar, Gregor Rozinaj, Puneet Kumar Yadav · 2022
Fast growth enhancement of web application has leaded to various security issues regarding intrusion not only in computer network framework but also web application themselves. Most of the techniques now in use in the context of Online Intrusion System (WIS) are unable to keep up with the complex and dynamic nature of cyber-attacks on web applications and issues linked to it. Attackers are using different techniques day by day to exploit the vulnerability of web applications. In order to identify intruder behaviors in order to predict future attacks on any web site, authors created an intrusion detection system (IDS) employing various machine learning and deep learning algorithms. The web application’s log files will be used to monitor any forthcoming requests, which will then be forwarded to our eight variously trained detection models. Anomaly detection model will use to identify unseen attacks also known as zero-day attack. The process of feature extraction and data processing is done for each different by itself to detect specific attack. After, experiments the evaluations shows that the average accuracy of all the models is 99.3% on benchmark data set–CSIC 2010 HTTP and ECML/PKDD 2007.