Publication Date

Fall 2015

Degree Type

Master's Project


Computer Science


Phishing is one of the prevalent techniques used by attackers to breach security and steal private and confidential information. It has compromised millions of users’ data. Blacklisting websites and heuristic-based methods are common approaches to detect a phishing website. The blacklist method suffers from a window of vulnerability. Many heuristics were proposed in the past. Some of them have better accuracy but a lower performance. A phishing filter should have better accuracy and peformance. It should be able to detect fresh phishing websites. Jo et al. [2] present a list of attributes of the web page to find the disparity between an original website and a spoofed website. The main aim of this project is to integrate the approach presented by Jo et al. [2] into web browser via Firefox add-on. Our phishing filter collects the list of attributes and compares it with the help of approximate string matching algorithms and WHOIS [14] server queries. For machine learning techniques, we used Weka [21]. All the algorithms, available in Weka were applied to our testing data set. Our phishing filter achieves 94.3% accuracy with reasonable performance.