A BioBERT-Based Intent Detection System for Personalized Doctor Recommendations

Abstract

In the medical field, clear and precise patient-reported symptoms are essential for ensuring appropriate medical advice from doctors. Because symptoms can be sent in a variety of languages, this method frequently encounters difficulties automatically interpreting the patient's purpose. This work suggests using the BioBERT (BiDirectional Encoder Representations from Transformers for Biomedical Text Mining) model, an NLP model tailored for biomedical text, to address this issue and increase the precision of identifying the purpose of patient medical symptoms. Data collection on symptoms , text preparation, intent recognition using the BioBERT model, and accuracy and loss measures are some of the techniques employed. The findings demonstrate that the BioBERT model may greatly increase detection accuracy when compared to conventional NLP models. This model achieves an accuracy of 94%, higher than previous approaches, with a lower loss value. These findings indicate that the use of BioBERT in text-based intent detection systems has great potential in improving the efficiency of technology-based medical services.