Detection and Analyzing Phishing Emails Using NLP Techniques

Abstract

The most common detrimental technique used by attackers to deceive victims into disclosing personal information is phishing, in which they pose as trustworthy individuals or organizations often via email. Although fake email attacks are a common tactic used by cybercriminals, their use has recently increased as attacker's profit from victims' anxiety. As a result, further study is required to determine how to recognize bogus emails. This paper proposed a new model to extract the Arabic email content and compare it using three determinants based on neural language programming (NLP) for the purpose of discovering whether it is a legitimate email or a phishing email. The first is a black list of Arabic common phishing words, the roots of a black list of Arabic common phishing words, and a list of Arabic common phishing sentences, the best two results for applying the above conditions were (99% Legal and 96% Phishing) when using the three conditions together and (99% Legal and 94% Phishing) when using a blacklist of common words of phishing, and then will present and discuss the results obtained.