Study of User Behavior on Phishing Mitigation of Risks

Abstract

In today s technological world, everyone is using Internet for daily works and banking newlinepurpose. Email has become standard way of communication and sharing information. newlineInternet security has always been a topic of discussion as there are new threads coming newlineeveryday by attackers. Phishing is one of the types of cyber-attack in which attacker uses newlinethe fake message and fake sites for trapping information money from clients. Attackers newlineattempt to draw online clients by persuading them to share their username, passwords, newlinebank account information for filling the billing data. One of the principal issues of phishing newlineemail location is the unknown zero-day phishing attack, which builds the level of trouble newlineto recognize phishing email. These days, phishers are making diverse portrayal strategies newlineto make unknown zero-day phishing email to break the barriers of those locators. There newlineare number of techniques are developed for phishing detection, but these systems were newlinefailed to provide appropriate results. Today, due to lack of current-age phishing detection newlinetechnique, phisher is managed to gain users credentials, and this causes financial loss of newlineusers. This work covers the study of existing phishing techniques and proposed technique newlinefor phishing detection. In this work, different machine learning techniques are used for newlinedeveloping phishing detection techniques. Machine learning algorithms used are J48, newlineNa¨and#305;ve Bayes, decision tree and multi-layer perceptron classifier. Proposed method shows newlineresult with more than 92% accuracy in detecting phishing websites using J48 classifier. newlineJ48 does best classification on spam base which is 97% for true positive and 0.025% false newlinenegative. Random forest works best on small dataset that is up to 5000 and number of newlinefeatures are 34. Na¨and#305;ve Bayes work faster for increased dataset size and reduced feature newlineset. newline newline