Article: An ensemble classification model for improved performance of phishing detection system Journal: International Journal of Information and Computer Security (IJICS) 2025 Vol.26 No.1/2 pp.68 - 90 Abstract: Individuals and organisations are at risk of money losses and data compromise from phishing attempts. Traditional rule-based phishing detection methods fail to keep up with attacker strategies. The need for more advanced and adaptive phishing defences is growing. An ensemble classification model for phishing detection system is proposed to address this difficulty. We analysed a large dataset of known phishing assaults to identify common patterns and traits. Initially, the raw phishing data have pre-processed using quantile transformation to convert variable's distribution to normal distribution and to mitigate the impact of outliers. The proposed ensemble prediction model has segregated phishing e-mails, webpages, and other dangerous information. The classifier's performance is enhanced further by employing the random search approach to tweak a set of carefully chosen hyper-parameters. When compared to other state-of-the-art methodologies, our system's results have achieved competitive performance. Thus, the suggested model is a promising ensemble-based phishing detection solution. Inderscience Publishers - linking academia, business and industry through research

Title: An ensemble classification model for improved performance of phishing detection system

Authors: Moumita Sahoo; Sabyasachi Samanta; Soumen Ghosh

Addresses: Haldia Institute of Technology, Hatiberia, ICARE Complex, Dist-Purba Medinipur, Haldia, West Bengal 721657, WB, India ' Haldia Institute of Technology, Hatiberia, ICARE Complex, Dist-Purba Medinipur, Haldia, West Bengal 721657, WB, India ' Haldia Institute of Technology, Hatiberia, ICARE Complex, Dist-Purba Medinipur, Haldia, West Bengal 721657, WB, India

Abstract: Individuals and organisations are at risk of money losses and data compromise from phishing attempts. Traditional rule-based phishing detection methods fail to keep up with attacker strategies. The need for more advanced and adaptive phishing defences is growing. An ensemble classification model for phishing detection system is proposed to address this difficulty. We analysed a large dataset of known phishing assaults to identify common patterns and traits. Initially, the raw phishing data have pre-processed using quantile transformation to convert variable's distribution to normal distribution and to mitigate the impact of outliers. The proposed ensemble prediction model has segregated phishing e-mails, webpages, and other dangerous information. The classifier's performance is enhanced further by employing the random search approach to tweak a set of carefully chosen hyper-parameters. When compared to other state-of-the-art methodologies, our system's results have achieved competitive performance. Thus, the suggested model is a promising ensemble-based phishing detection solution.

Keywords: cyber security; phishing; data breaches; quantile transformation; random forest ensemble classifier; RFEC; hyper-parameter tuning; cross-fold validation.

DOI: 10.1504/IJICS.2025.145112

International Journal of Information and Computer Security, 2025 Vol.26 No.1/2, pp.68 - 90

Received: 10 Sep 2023
Accepted: 23 Mar 2024
Published online: 19 Mar 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: An ensemble classification model for improved performance of phishing detection system

Keep up-to-date