Enhancing The Cyber Security Using Ensemble Stacking Model For Phishing Sites Detection With Hyperparameter Tuning
Md. Mahedi Hassan, Ahsan Ullah, Anup Chakraborty, Nurunnabi Sarker, Bikash Kumar Saha Roy · 2024
Phishing attacks have become a significant cybercrime in recent times. Phishing is a cyber attack that tricks online users into revealing sensitive information through a fake website that looks legitimate. Attackers with stolen credentials with stolen credentials not only use them for the targeted website but also use them to access other popular legitimate websites. There are many anti-phishing techniques, toolbars, and extensions, yet phishing attacks remain a major concern in the digital world. The proposed study is based on the phishing URL-based dataset extracted from the UCI machine learning repository, which consists of 11000+ instances and 30 features. After applying various preprocessing and feature selection techniques, many machine learning algorithms have been applied and designed to prevent phishing URLs and provide protection to the user. This study uses a stacking ensemble, which is a combination of different types of machine learning models such as decision tree (DT), random forest (RF), naive Bayes (NB), gradient boosting classifier (GB), K-neighbors classifier (KNN), support vector machine classifier (SVM), Ada boost classifier (ADA), LightGBM Classifier (LGBM), XGBoost Classifier (XGB), Multi-layer Perceptron classifier (MLP) to defend against phishing attacks with high accuracy and efficiency. For meta-learners, we used Logistic Regression (LR). To enhance the performance of our proposed model, we applied random search hyperparameter tuning. To guarantee more reliable results, and explainable artificial intelligence (XAI) technologies at 97.40% accuracy, the best result was produced by our proposed model. The comparative analyses’ results demonstrate that the proposed approach outperforms the other models and achieves the best results.