Detection of Web-Attack using DistilBERT, RNN, and LSTM
Biodoumoye George Bokolo, Lei Chen, Qingzhong Liu
Abstract
The rise in usage of the Internet has tremendously helped those who use web applications. Web-based applications are becoming more susceptible to numerous security risks and network vulnerabilities as online attacks continue to develop. Malicious code or contents could be embedded in requests from HTTP causing attacks like SQL injections etc.In this research, an online intrusion detection system is presented to tackle the rise in web application attacks. Our web intrusion detection system uses a Distil-BERT, RNN, and LSTM model to identify attacks with body, URL, and User-data. The experimental findings demonstrate that our model successfully classifies the attacks with body, URL, and user data with a 94% accuracy.