Intent and URL based ML Classification for Phishing Detection

Reports show that the number of phishing web sites is exponentially increasing and it is estimated that between 80% to 93 % of the data breaches are involving phishing attacks. With both probability of occurrence as well as impact high, it is a high risk calling for an effective solution. Linear approaches like multi-factor authentication currently practiced addressing an exponentially growing problem is inadequate. In this work, we present a solution using domain-based understanding of the attacker intent based on the attributes of information that the attacker ultimately seeks to know from the user and make it more effective by combining it with machine learning based solutions of phishing detection based on URL features.