A TRANSFORMER-BASED DISTILBERT FRAMEWORK FOR REAL-TIME DETECTION OF PHISHING ATTACKS IN SMS AND WHATSAPP

The rapid growth of mobile communication platforms such as Short Message Service (SMS) and WhatsApp has significantly increased the threat of phishing attacks, leading to severe financial losses and privacy breaches.Traditional phishing detection approaches based on handcrafted features and classical machine learning models often fail to capture the con-textual semantics of short text messages.To address this limitation, this paper proposes a transformer-based DistilBERT framework for real-time detection of phishing attacks in SMS and WhatsApp messages.The DistilBERT model is fine-tuned on the publicly avail-able SMS Spam Collection dataset after appropriate text preprocessing and tokenization.The performance of the proposed framework is evaluated using standard classification metrics in-cluding accuracy, precision, recall, and F1-score.Experimental results show that the proposed model achieves an accuracy of 99.28%, a precision of 97.96%, a recall of 96.64%, and an F1score of 97.30%, demonstrating outstanding phishing detection capability and strong gen-eralization performance.Owing to its lightweight architecture and high detection accuracy, the proposed framework is well suited for real-time deployment in mobile security applications.