Phishing URL Detection Using RNN - LSTM Models for Safer Web Browsing

Document Type

Conference Proceeding

Source of Publication

2026 IEEE 5th International Conference on AI in Cybersecurity Icaic 2026

Publication Date

2-18-2026

Abstract

This position paper was written by Master students of Zayed University with the aim of introducing a new RNN-LSTM phishing URL detector. The paper will be a systematic literature review of recent IEEE and ACM research to provide an analysis of the strengths and limitations as well as the comparative performance of traditional machine learning, independent deep learning, and hybrid architectures in URL based phishing detection. The review points out the key significant weaknesses of existing models, such as being computationally expensive, uninterpretable, and difficult to deploy in real time on resource constrained platforms such as browsers and mobile phones. These findings are extrapolated by the paper in determining the gap in research on the lightweight, adaptive detection frameworks able to strike a balance between accuracy and efficiency to be practically deployed in browsers and mobile phones. The paper does not provide experimental findings made in a model; rather, it sets out a conceptual framework, explains the research gaps, and justifies why it is necessary to conduct additional experimentation. This research thus provides the groundwork in future experimental research and testing of a lightweight, adaptive phishing recognition model that could be deployed in a resource limited client environment.

ISBN

[9781665477611]

Publisher

IEEE

Disciplines

Computer Sciences

Keywords

Browser security, Character-level, cybersecurity, Deep Learning, Lightweight Models, Machine Learning, Phishing URLs, Real Time Detection, RNN+LSTM, social engineering, URL Detection

Scopus ID

105041771809

Indexed in Scopus

yes

Open Access

no

Link to Full Text

Share

COinS