Skip to content
Conference

Intelligent Website Phishing Detection Based on a Comparative Study of Ensemble Supervised Learning Techniques

Sep 2026 · 2026 International Conference on Data Science and Its Applications (ICoDSA) · pp. 607-612 · 0 citations · 21 references

Abstract

Phishing websites is one of the methods used by the attacker to steal sensitive information from the users. Traditional phishing detection methods such as signature based and rule-based techniques often fall short in adapting to newly emerging phishing strategies, resulting in high rates of false positives and false negatives. This research compared Voting and Stacking ensemble learning approaches using experiments of detecting phishing websites. The dataset used in this experiment contained 88647 URLs and 111 extracted features. K-Nearest Neighbors, Support Vector Machine and Logistic Regression are the models used for the ensemble learning approaches. The machine learning models were tuned by Hyperparameter tuning using the method of GridSearchCV. The machine learning models were evaluated using K-Fold Cross Validation. The performance of all the models were measured using accuracy, precision, recall, F1-score, ROC AUC and PR AUC. After the experiment has executed, the results shows that Stacking ensemble learning approaches are the best performer with 96.42% accuracy, 95.18% precision, 94.80% F1-score, 99.29% ROC AUC and 98.73% PR AUC. The contribution of this study lies in highlighting the limitations of static detection techniques and promoting the application of intelligent ensemble models for phishing website detection. From a practical perspective, the outcomes of this research can assist cybersecurity practitioners and organizations in deploying more accurate and adaptable phishing detection systems that can better protect users against online threats.

View source

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.