A Novel DEA-ELM Hybrid Method for Web Phishing Detection

OpenAlex İngilizce

Phishing attacks are a pervasive cybersecurity threat, using deceptive web pages to steal users' sensitive information. Detecting phishing sites with high precision and efficiency is crucial for building effective countermeasures. In this study, we propose a novel classification model that integrates a Differential Evolution Algorithms (DEA) with Extreme Learning Machines (ELM) framework for phishing website detection. The approach introduces a DEA mechanism for inter-feature signal enhancement and couples it with an ELM, optimized through a DEA. The proposed DEA-ELM model was evaluated on the Web Page Phishing Detection dataset, Compared to traditional machine learning models such as Random Forest, Logistic Regression, Support Vector Machine (SVM), and Decision Tree, which achieved accuracies between 93% and 97%, the proposed DEA-ELM model achieved a remarkable 99.86% accuracy, along with high precision, recall, and F1-score metrics. These results confirm the potential of DEA-optimized ELM combined with DEA analysis in creating scalable, accurate, and real-time phishing detection systems. The model also provides a reproducible framework by using publicly available data and open-source feature extraction scripts. Future work may explore hybrid feature selection strategies, larger-scale deployment, and online learning extensions.