Phishing email detection using learning vector quantization

Abstract

Phishing attack is defined as the attempt to get valuable information such as password, credit card information by gaining the trust of the users in electronic communication. Phishing detection system could possibly protect many losses that are happening due to the phishing attack. A background research performed during the project showed that today internet world faces many security threats where one of the major security threat is phishing. Phishing attacks create a serious risk for end-users. A thorough survey of approaches of defense mechanisms for the detection of the phishing e-mail attack have been discussed. Phishing attacks are growing in great number every day; hence it becomes vital for us to take a stance to defend the email users from phishing attack. The proposed phishing email detection system will detect phishing mails using a technique known as Learning Vector Quantization. This phishing detection system will retrieve the unread emails of the users, classify the emails, detect the attack and alert the user about the attack. In this project, the phishing detection system was designed, developed and tested. Datasets were developed to train and test the models. Finally, the technique used in the application was evaluated against other similar techniques to determine the effectiveness of the system which proves that the proposed technique achieved higher accuracy and lower false positive and the future works have been suggested. v Table of Contents Abstract