Performance exploration of random forest (RF) and naive bayes (NB) classifiers in identification of phishing attacks

Abstract

With the expansion of the web, more and more programs are being hosted online and accessible in this way. Because of this change, a hacker has begun targeting computers using phishing websites. To identify a phishing attempt, many methods have been offered. However, more work has to be done to counteract this phishing risk. The purpose of this research is to examine and assess the machine learning approach’s performance in classifying phishing attacks. In order to detect phishing attempts inside the web site applications, this work used a heuristic strategy using machine learning classifier. This research evaluates three different machine learning classifiers for their ability to detect phishing scams, contrasting them with random forest (RF) and Naive bayes (NB). It shows that Random forest can successfully detect phishing assaults with a true positive rate of 96.65% by making use of characteristics unique to each website. Based on the outcomes, it seems to be a reliable classifier for identifying phishing scams.