Novel Optimization-Driven Feature Selection Approach for Enhancing Phishing Website Detection Accuracy

Abstract

Phishing attacks continue to pose significant threats to cybersecurity, prompting the need for robust detection mechanisms. This study introduces an optimization-driven feature selection approach aimed at enhancing the accuracy of phishing website detection. By systematically selecting and prioritizing relevant features, guided by advanced optimization techniques, the proposed approach outperforms traditional baseline methods across various evaluation metrics. Experimental results demonstrate notable improvements in detection accuracy and robustness compared to conventional techniques. The optimization-driven approach offers scalability, adaptability, and efficiency in navigating the feature space, making it suitable for diverse datasets and scenarios. This study contributes to the advancement of phishing detection systems by providing a systematic and effective approach for identifying relevant features and minimizing false alarms and false negatives. Future research should explore the integration of machine learning models and conduct real-world validation studies to further validate the effectiveness and generalizability of the proposed approach.