Two-level authentication approach to protect from phishing attacks in real time

Abstract

Nowadays, the phishing attack is emerging as serious Internet security threat, which causes massive financial losses every year. There are various approaches available to detect phishing attack, e.g., blacklist, machine learning, visual similarity, etc. However, most of these approaches have various limitations, as they are complicated, produce high false positive rate, language dependent, slow in nature, and not fit for the real-time environment. In this paper, we present a two-level authentication approach, which not only detects phishing attacks accurately in real time environment but also does not depend on the textual language of the webpage. Proposed approach execute two authentications before declaring a webpage as phishing, which makes it more accurate, reliable, and fast. In the first level authentication, the search engine based mechanism is proposed which use a simple, reliable and language independent query to authenticate the webpage. The second level authentication processes different hyperlinks within the source code of the webpage for the detection of phishing webpages. Performance of the proposed approach is evaluated, and it achieved the significantly higher true negative rate of 99.95%. Comparison with other existing methods also proves the supremacy of our proposed approach.