Phishingwebpage detection usingweighted URL tokens for identity keywords retrieval

Abstract

Phishing is an online identity theft that has threatened Internet users for more than a decade. This paper proposes an anti-phishing technique based on a weighted URL tokens system, which extracts identity keywords from a query webpage. Using the identity keywords as search terms, a search engine is invoked to pinpoint the target domain name, which can be used to determine the legitimacy of the query webpage. Experiments were conducted over 1000 datasets, where 99.20% true positives and 92.20% true negatives were achieved. Results suggest that the proposed system can detect phishing webpages effectively without using conventional language-dependent keywords extraction algorithms.