"DAKSHAAB": An AI-Driven Chatbot for Patient Triage Using Bio-Bert

The specialist's referral in the treatment of patients is one of the essential measures of raising the quality of the patient's care. Structures depend upon typical styles and manners which not only limit the accuracy of the matching of specialists but their scope of domain as well. This paper develops a novel solution to enhancing the process of referrals, specifically a novel chatbot powered by Bio-BERT – one of the most sophisticated models for understanding biomedical language – using other advanced NLP techniques. Our system is trained using Bio-BERT on internal data comprised of Bio and review comments of patients to ensure the chatbot considers minute details in the complaint and yearning of the patient. Certain pre- processing techniques as applied to healthcare literature biosystems and textual data greatly enhance the accuracy and relevance of specialist recommendations. Accuracy, precision, recall, and F1-score are among the metrics subjected to various sets of evaluation, where a near improvement is observed with accuracy always being above 99.7% at each epoch. Such results endorse Bio-BERT's accuracy at managing the intricacies of medical language for a conceptually flawless patient specialist matching service. The disruption these technologies present in improving specialist referral systems is particularly underlined in this study.