WEB PHISH DETECTION (AN EVOLUTIONARY APPROACH)

Amruta Deshmukh, Sachin Mahabale, Kalyani Ghanwat, Asiya Sayyad

Abstract: Phishing is nothing but one of the kinds of network crimes. This paper presents an efficient approach for detecting phishing web documents based on learning from a large number of phishing webs. Phishing means to make something fraud with someone, usually by using internet with the help of emails, to take our personal information, such as credentials. The finest way to protect ourselves and our credentials from phishing attack is to understand the concept of phishing as well as to understand that how to determine a phishing attack. Most of the phishing emails are sent from well-reputed organizations and they ask for your credentials such as credit card number, account number, social security number and passwords of bank account. Mostly the phishing attacks seen from the websites, services and organizations with which we do not even have an account. In this system we are using two classifiers to detect phishing. To recognize the phishing, the Uniform Resource Locator (URL) features of the website are firstly analyzed and then they are classified by using K-means classifier. If the answer is still suspicious then by using parsing of the webpage, its DOM tree is drawn and then the second classifier that is Naive Bayesian (NB) classifier classifies the web page.

Keywords: phishing, phishing emails, classifier

DOI: https://doi.org/10.15623/ijret.2014.0301095