Intelligent System for Phishing detection on web pages using Random Forest (#971)

Authors

Araujo Arévalo, Álvaro Sebastián

Felix-Diaz monzon, Gonzalo Amed

Mansilla López, Juan Pablo Jesús

Abstract

During the last years there have been many technological problems, among them, one of those that stands out for its impact worldwide is Phishing, which is a Social Engineering technique that seeks as a benefit to illegally appropriate confidential and private information. In Peru, this problem is spreading considerably due to the lack of interest of both state and private entities in being able to make a contingency plan against this problem, and more and more citizens are affected and there is no clear statement or response from the authorities. In response to this major problem, this paper proposes the use of an intelligent system that uses a Random Forest algorithm, which is based on machine learning and can detect in time the URLs that may be malicious. The operation of the system is based mainly on the analysis of the characteristics of the URLs, which is complemented with a dataset with which the algorithm was trained, and thus give a comprehensive detail of analysis. According to the surveys made to the users of the system, a great acceptance of the intelligent system was obtained due to the precision that it has and, in this way, avoid falling in the deceptions based on phishing techniques. The results of the use of the Random Forest algorithm for this project were good, having an accuracy of 0.98 compared to 0.93 of the Decision Tree algorithms and 0.91 of Neural Networks algorithm.