Shark-Eyes: A multimodal fusion framework for multi-view-based phishing website detection
Minh Vo Quang, Bui Tan Hai Dang, Ngan Tran Kim Ngoc, Ngo Duc Hoang Son, Nguyen Huu Quyen, Phan The Duy, Van-Hau Pham
Abstract
In the era of escalating cyber threats, phishing attacks continue to exploit vulnerabilities in online security. This paper presents Shark-Eyes, a novel multimodal fusion framework designed for the detection of phishing websites using a multi-view approach. The proposed approach leverages a combination of two distinct attributes, namely domain features and HTML tag features, extracted from the target websites. The framework’s effectiveness is evaluated through comprehensive experiments on a dataset sourced from Phishtank, OpenPhish, and Alexa, encompassing real-world phishing instances. Our results demonstrate the robustness and efficiency of the Shark-Eyes framework in accurately identifying phishing websites, showcasing its potential as a powerful tool for enhancing online security and thwarting malicious activities.