A Detailed Analysis on Various Datasets using Machine learning and Deep Learning Techniques for Phishing URLs Detection

Abstract

Phishing represents a deceitful interaction where an assailant attempts to acquire delicate data from the person in question. Phishing is a sort of digital assault, where the attacker takes the individual data of the user or victim and uses this information for malicious purposes. To resolve this problem many techniques are there but each technique has some advantages and the wall has some disadvantages. Phishing attack is increasing day by day so it's a more critical problem which faces by each and everybody who used the internet for various purposes. Phishing is a bogus cycle, where an aggressor attempts to get significant data from the person in question. By and large, these sorts of assaults are performing utilizing instant messages, messages and sites. Noxious sites, which are right now in a broad ascent, have the same look as real destinations. Alternately, their backend is intended to gather astute data that is inputted by the harmed party. Nonetheless, their backend is intended to gather delicate data that is inputted by the person in question. In this paper, we selected the dataset based on the performance of various ML and DL models. Also compare the accuracy of these models with existing work and select the best features of URLs to detect phishing attacks.