URL-based phishing detection using hybrid ensemble technique
Phishing attacks pose a significant and growing threat to cybersecurity by deceiving users into disclosing sensitive information through malicious websites. Most existing URL-based phishing detection studies rely on individual classifiers or traditional ensemble techniques, which often struggle to...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | en |
| Published: |
Penerbit Universiti Teknikal Malaysia Melaka
2025
|
| Online Access: | http://eprints.utem.edu.my/id/eprint/29512/2/02773020120261136232879.pdf http://eprints.utem.edu.my/id/eprint/29512/ https://jacta.utem.edu.my/jacta/article/view/5322 https://doi.org/10.54554/jacta.2025.07.02.003 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Phishing attacks pose a significant and growing threat to
cybersecurity by deceiving users into disclosing sensitive information through malicious websites. Most existing URL-based phishing detection studies rely on individual classifiers or traditional ensemble techniques, which often struggle to generalize against evolving phishing patterns. To overcome this limitation, this study proposes a hybrid ensemble learning approach for UR based phishing detection by integrating a Random Forest Classifier with AdaBoost, Bagging, and Stacking strategies. Experiments were conducted using a publicly available benchmark dataset from the UCI Machine Learning Repository consisting of 11,055 URLs and 30 features. Model performance was evaluated using 10-fold cross-validation. The results show that the Random Forest–Stacking hybrid model achieved the highest accuracy of 97.21%, outperforming other hybrid configurations. The findings demonstrate that stacking-based hybrid ensemble learning enhances generalization and
robustness in phishing URL detection. |
|---|