A novel phishing detection system using binary modified equilibrium optimizer for feature selection

doi:10.1016/j.compeleceng.2022.107689

A novel phishing detection system using binary modified equilibrium optimizer for feature selection

2022-03-01

20

PRE

AI

DOI：10.1016/j.compeleceng.2022.107689

Original

Original request for help

Save

Abstract

En 中文

The Digital era faces security concerns due to the explosive growth of cyber-attacks, such as phishing attacks, man-in-the-middle attacks, and many more. Phishing attackers deceive users by legitimate counterfeit websites that tend users to provide confidential information on a phishing website. This paper proposes a new phishing detection system that uses Binary Modified Equilibrium Optimizer (BMEO) with a proposed AV-shape transfer function (AV-BMEO) and k-nearest neighbor classifier. AV-BMEO having high exploration and exploitation abilities is used for the feature selection and classifier's hyperparameter optimization. The AV-shape transfer function is designed based on opposition-based learning to enhance the proposed algorithm's exploration capabilities. The statistical validation proves the better performance of AV-BMEO compared to seventeen algorithms in terms of accuracy and number of selected features on eighteen datasets. Further validation shows an improved prediction of phishing websites with the optimum number of features compared to several state-of-art techniques on phishing datasets.

Keywords:

Phishing

AV transfer function

Modified equilibrium optimizer

Feature Selection

Metaheuristic

A novel phishing detection system using binary modified equilibrium optimizer for feature selection

Abstract

Journal

Organization