An Empirical Analysis of Machine Learning Techniques in Phishing E-mail detection
Arriane Livara, Rowell Hernandez
Abstract
Communication through email is the most secure medium to transfer files, sensitive information, and messages in this new normal where most people work from home and classes are already online. Attackers take advantage of this by implementing social engineering attacks, and the most commonly known attack is Email Phishing. This attack aims to lure the individual into clicking malicious links that can automatically steal personal information like bank and credit card accounts, client details, passwords, and more. This research aims to develop multiple machine learning models to classify if the email is legitimate or a phishing email and recommend the most suitable model that performed the best and discuss how accurately it performed. The classifiers used achieved 99% of accuracy for detecting phishing and legitimate emails even though trained with an unbalanced data.