Machine Learning Approaches for Proactive Phishing Attack Detection
B. T. Geetha, P. Malathi, T. Thirumalaikumari, V. Janakiraman, H. Anwer Basha, S. Rukmani Devi · 2024
Phishing attacks remain a serious problem in cybersecurity, necessitating advances in detection methods. The study describes a novel method for detecting phishing attacks that use advanced machine learning (ML) techniques, notably convolutional neural networks (CNNs) and recurrent neural networks (RNNs). CNNs and RNNs were chosen because of their ability to detect spatial and sequential patterns in phishing data. The system achieves improved performance by iterative model refinement and ensemble learning, beating existing rule-based systems in terms of precision (96%), recall (94%), and F1-score (95%). Furthermore, the proposed strategy has significantly lower false positive (3.8%) and false negative (5.7%) rates than existing systems, indicating greater dependability in distinguishing between legitimate and malicious emails. Cross-validation validates that the proposed system is stable and ready for real-world use. As a result, the proposed ML-driven strategy represents a substantial leap in phishing detection, providing increased accuracy (F1-score: 95%) and resilience to emerging cybersecurity threats.