Securereg: Combining Nlp and Mlp for Enhanced Detection of Malicious Domain Name Registrations

Abstract

The escalating landscape of cyber threats, charac-terized by the registration of thousands of new domains daily for lar ge-scale Inter net attacks such as spam, phishing, and drive-by downloads, underscor es the imperati ve for innovative detection methodologies. This paper introduces a cutting-edge approach for identifying suspicious domains at the onset of the registration process. The accompanying data pipeline generates crucial featur es by comparing new domains to register ed do-mains, emphasizing the crucial similarity score. The proposed system analyzes semantic and numerical attrib utes by leveraging a novel combination of Natural Language Processing (NLP) techniques, including a pretrained CANINE model and Multilayer Perceptr on (MLP) models, providing a robust solution for early threat detection. This integrated Pretrained NLP (CANINE) + MLP model showcases the outstanding perf ormance, surpassing both individual pretrained NLP models and standalone MLP models. With an PI score of 84.86% and an accuracy of 84.95%on the SecureReg dataset, it effecti vely detects malicious domain registrations. The finding demonstrate the effecti veness of the integrated appr oach and contrib ute to the ongoing efforts to develop proactive strategies to mitigate the risks associated with illicit online activities through the ear ly identificatio of suspicious domain registrations. © 2024 IEEE.