Phishing detection is a critical area in cybersecurity that significantly impacts the protection of sensitive information and the overall security posture of individuals and organizations. The increasing sophistication of phishing attacks presents substantial challenges in identifying fraudulent websites that impersonate legitimate entities. Existing detection methods often struggle with high false positive rates and misclassification errors, highlighting the need for more effective solutions. In response to these challenges, this study developed and evaluated a Multi-Layer Perceptron (MLP) model specifically designed for phishing website detection. The research utilized a comprehensive dataset containing features extracted from both legitimate and phishing websites, combining textual and numerical attributes to enhance classification performance. The MLP model was rigorously assessed using metrics such as accuracy, precision, recall, F1- score, and the area under the ROC curve (AUC). Results indicate an overall accuracy of 96.6%, a precision of 96.5%, and a recall of 97.5%, along with an AUC of 0.9941. These findings showcase the model's strong discriminatory power and effectiveness in minimizing misclassifications. The research highlights a significant advancement in phishing detection capabilities compared to existing approaches, laying the groundwork for future developments in phishing detection systems. The study emphasizes the potential for real-world applications in enhancing cybersecurity defenses against evolving threats.
Ochu et al. (Wed,) studied this question.