Enhancement of Fake Website Detection Techniques Using Feature Selection and Filtering Algorithms

Hetal Rahul Rajpura, Hiteishi Diwanji


Phishing websites are malicious websites created by fraud people to mimic real websites. Phishing websites impersonates
legitimate websites to lure users into visiting the fake websites.They may lure by mailing you suspicious links that appear to be
legitimate.Phishing is a security threat to the Internet, which causes tremendous loss every year. It helps generate billions of revenue for
fraudsters. Attackers may steal user private information, credit-card and other significant financial information. In this paper , a study on
various phishing websites detection techniques such as WHOIS, Browser-Integrated Anti-phishing toolbars,attaching new top level
domains to existing blacklisted urls, feature extraction and classification based approach, a hybrid phish detection method based on
information extraction and information retrieval, CANTINA, a content-based approach, purely based on the TF-IDF(term frequency
/inverse document frequency) used in information retrieval algorithm , a novel classification method that identifies malicious web pages
based on static attributes is presented.On detailed study and analysis,it is found that there exists classification approach based on the
decision tree algorithm, which has mean absolute error of 0.292 which is comparatively larger than other algorithms, whereas its prediction
accuracy is 98.5%. Hence, improvements can be done to decrease the mean absolute error of decision tree by using feature selection and
filtering techniques to classify a revised set of attributes and design a sound and a
robust phishing detection system .

Keywords: Phishing, Detection, Techniques, Classifiers, Feature reduction, Filters, Security

Full Text:


DOI: https://doi.org/10.26483/ijarcs.v4i3.1551


  • There are currently no refbacks.

Copyright (c) 2016 International Journal of Advanced Research in Computer Science