Machine Learning based Phishing Website Detection

Abstract

Phishing is a technique to obtain or exploit the personal information of an individual by imitating an existing website or by offering interesting schemes through email, text messages. Phishes steal important and secured information like passwords, credit card details, phone numbers. Nowadays phishing attacks are increasing which is extremely problematic for social and economic websites. The prime focus of the paper is to build a powerful application that applies Machine Learning techniques and tools to identify phishing websites. Training with one classification model is not the best way in the case of predicting websites because accuracy plays an important role. Therefore, we consider various Machine Learning algorithms such as Random Forest(RF),Logistic Regression model(LR), Support Vector Machine(SVM) or maximum-margin classifier, Decision Tree(DT), Sequential Multilayer Perceptron(MLP), Naïve Bayes(NB).  After reviewing each algorithm we select a classification model with the highest accuracy to detect new fake websites given by the user.