Big Data Security Analytics for Phishing URLs Detection

Abstract

Phishing is a type of cybercrime in which uninformed internet users are duped into disclosing personal information such as login credentials and credit card information. Unlike software vulnerabilities, phishing attempts target human weaknesses and can be difficult to detect. We created multiple machine learning models in this research to detect phishing URLs based on their attributes. The models were trained in three ways: with all features, with feature selection, and with feature reduction. We employed approaches such as principal component analysis, ensemble modelling, mutual information gain, and stacking development to increase the performance of the models.