Full Program
Summary:
In this work, we present multiple techniques for detecting fraudulent web pages, utilizing our own dataset. We developed two primary approaches: a URL-based method and a content-based method, which were later combined to enhance overall performance. Both approaches leverage a combination of hand-engineered features and automatic feature extraction, incorporating advanced NLP methods and large language transformer models. Our study emphasizes not only the accuracy of phishing website detection but also the efficiency of the models in terms of inference time.Author(s):
Andreea Dance
Technical University of Cluj-Napoca
Romania
Andreas Fraticiu
Technical University of Cluj-Napoca
Romania
Ciprian Oprisa
Bitdefender
Romania