Web application firewall using deep learning algorithm

Mohd Nawi, Mohd Amir Faris (2023) Web application firewall using deep learning algorithm. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) Web application firewall using deep learning algorithm.pdf - Submitted Version Download (482kB)
	Text (Full text) Web application firewall using deep learning algorithm.pdf - Submitted Version Restricted to Registered users only Download (7MB)

Abstract

Web application firewalls (WAFs) are essential for defending web applications against a variety of online dangers. Traditional WAFs, however, frequently find it difficult to keep up with the continually changing threat scenario. Deep learning techniques have been a viable solution to this problem for improving the performance of WAFs. In-depth research on the use of deep learning in web application firewalls is presented in this publication. This research explore the fundamentals of deep learning and its relevance in the context of cybersecurity and WAFs. A detailed analysis of different deep learning architectures, such as Convolutional Neural Networks (CNNs), Long Term Short Memory (LSTM), and support vector machines (SVM), is presented to understand their potential contributions to the WAF's defensive capabilities. Data preprocessing, feature extraction, and model training are all part of the implementation phase for integrating deep learning models into the WAF. Given that web application attack datasets are frequently skewed toward regular traffic, addressing imbalanced datasets is given special consideration. In order to increase model generalization, a number of ways to supplement the sparse labeled data are investigated. The results of this study show that integrating deep learning into WAFs can greatly improve their security capabilities by efficiently identifying and thwarting a variety of web application assaults. The outcomes also emphasize the necessity of ongoing model modification and adaptation to maintain resistance to new dangers. Deep learning is an appealing solution for the changing landscape of web-based cyber threats even if it adds more computational cost and has performance trade-offs.

Actions (login required)

View Item