A Review on Detecting Phishing URLs using Clustering Algorithms

Phishing is a kind of a social engineering attack. The attacker poses as a legitimate entity and communicates with the victim through some mode of communication. The user is prompted to open a link which has been designed to look similar to a legitimate website, or is prompted to relay sensitive credentials over the phone. The attacker steals the users' information to perform identity theft, account hijacking, etc. In this paper, we focus on URL based phishing attacks. Most of the solutions that we investigated focused more on the classification algorithms rather than clustering. Our aim is to experiment and compare the results of both of these types of algorithms. The main premise of our approach is a hybrid machine learning model comprising of two steps- checking with a blacklist and whitelist, and heuristics based detection, to increase the accuracy of the proposed algorithm.