A Unified Approach on Phishing Detection using Chrome Extension Integrating ML and NLP

Phishing attacks have become a persistent and evolving threat in the digital landscape, exploiting human and system vulnerabilities to compromise sensitive data. Traditional phishing detection methods, such as blacklists and rule-based filtering, are increasingly ineffective due to their static nature and inability to adapt to novel attack vectors. This research proposes a unified approach to phishing detection that integrates Machine Learning (ML) and Natural Language Processing (NLP) to enhance detection capabilities and mitigate the limitations of conventional methods. By leveraging ML algorithms such as logistic regression and multinomial naive Bayes, the proposed system identifies structural and behavioral patterns indicative of phishing attempts. Concurrently, NLP techniques, including tokenization and stemming, are employed to analyze textual features within URLs and phishing messages, extracting semantic and syntactic characteristics crucial for classification. To validate the effectiveness of this approach, extensive experiments were conducted on real-world datasets containing labeled phishing and legitimate URLs. The results demonstrate that the integration of ML and NLP significantly improves classification accuracy while reducing false positive rates compared to traditional detection mechanisms. The combined approach enables adaptive learning, allowing the system to recognize emerging phishing tactics with minimal manual intervention. Feature engineering and model selection were found to be critical in optimizing detection performance, as careful selection of input features enhances predictive accuracy and robustness. These findings underscore the potential of ML and NLP in developing more resilient cybersecurity solutions, paving the way for advanced, real-time phishing detection systems.