HyRANN-UPD: Enhancing Phishing URL Detection Using Ridge Regression-Based Feature Selection and Artificial Neural Networks

Phishing attacks have become a major cybersecurity threat, making it essential to develop advanced detection models to protect online users.This study presents a machine learningbased approach for detecting phishing URLs, utilizing an Artificial Neural Network (ANN) to improve accuracy and reliability.The PhiUSIIL Phishing URL Dataset from the UCI Machine Learning Repository, containing 235,795 instances with 55 features, was used for training and evaluation.The dataset includes 134,850 legitimate URLs and 100,945 phishing URLs, with no missing values.To enhance performance, Ridge Regression (L2 Regularization) was applied to reduce the feature set from 55 to 50, improving efficiency without compromising accuracy.Several machine learning models which include Random Forest (RF), Nave Bayes (NB), Logistic Regression, K-NN, XGBoost, and ANN were tested to compare their effectiveness.Among them, the ANN model outperformed the others, achieving an accuracy of 98.58%, precision of 97.80%, recall of 97.66%, and an F1score of 97.65%.The ROC-AUC score of 0.98 further demonstrated the model's ability to differentiate between phishing and legitimate URLs.The proposed ANN model is efficient, scalable, and suitable for integration into existing security frameworks such as intrusion detection systems and anti-phishing tools.This research contributes to the growing field of AI-driven cybersecurity solutions, offering a highly effective and reliable approach to counter phishing threats.