Detecting phishing e-mails using text mining and features analysis

Phishing e-mails are used by malicious actors with the aim of obtaining sensitive information from a victim, deceiving or blackmailing them. An inattentive or uninformed user may often fail to recognise if an e-mail is sent by an authentic sender or is a scam. We therefore sought to develop a method that can effectively and efficiently detect phishing e-mails and report them to the user. We analyse all the information available on receipt of the e-mail both statically and performing text mining on the content and subject of the e-mail. In addition to indicating weather e-mails are suspicious, the degree of accuracy with which the above statement is made is also reported, and the aspects of the e-mail that are characteristic of a phishing e-mail are highlighted. Excellent results were achieved with our methodology, reaching 99.2% accuracy.