Email classification tool to detect phishing using hybrid features

Abstract

Phishing is a fraudulent attempt of trying to gather personal sensitive information such as user ID and passwords, credit card and bank account details through network. Social messaging and websites are used as medium to trigger attacks in addition to the use of emails, which is the most common and leading method currently used to perform phishing attacks. In an attack, the attacker is sending an email with a URL of the phishing website camouflaged as a legitimate source. Nowadays phishing has become more complicated and critical problem to many organizations. The phishers can bypass the filters and rules set by anti-phishing procedures and techniques. This research build a web based phishing email detection tool using data mining classification model. To build an efficient classification model, varieties of extracted email features have been used. These selected features can be categorized according to email header, email body, URL and Web Page Content of URL. In this model, classification accuracy will be enhanced by using these hybrid features. This model will be used to implement the web-based tool to detect phishing emails with more accuracy even without opening the emails. This can be used as preventive and proactive technique for phishing detection.