Phish-armour: phishing detection using deep recurrent neural networks

Abstract

Phishing is an illegal cybercrime, wherein a target gets victimized for sacrificing their personal and corporate information. It is one of the most straightforward forms of cyber-attack for hackers, as well as one of the simplest for victims to fall for. It can also provide hackers with the required information that are needed to access their targets’ personal and corporate accounts. For the past decade, machine-learning techniques have become consistent standards for classifying phishing and legitimate URLs. But deep learning algorithms have the advantage of automatic extraction of complex features and characterization of handling massive data. Considering the above-listed advantages, this work provides state-of-the-art accuracy in the detection of malicious URLs using recurrent neural networks (RNN). Unlike previous studies, which looked at online content, URLs, and traffic numbers, this work aims to focus only on the text in the URL which makes it quicker, and thereby zero-day assaults could be caught at the earliest. The RNN has been optimized so that it might be utilized on tiny devices like Mobiles, and Raspberry Pi without sacrificing the inference time.