Year: 2025
Type: article
Abstract: NLP are vastly changing the health sector in allowing intelligent systems to answer complex medical questions.This work is focused on developing chat and voice bots, which will use advanced techniques of NLP entailing cardio and ENT.By running machine learning algorithms on large clinical datasets, we have been able to construct a robust framework whose aim would be to increase the engagement of patients and provide relevant, accurate, and timely medical information for the same.These bots have been trained on a broad database of Cardio and ENT-related questions, with intensive preprocessing steps for quality and consistency assurance, including tokenization and lemmatization.Models refit using neural networks, especially recurrent and convolutional neural networks, to master the context of medical questions and hence understand them better, increasing the accuracy of the responses.Our experiments showed an accuracy of 90% in both text and voice inputs, regarding comprehension and appropriate response by the bots.A qualitative analysis of the bot responses showed a high relevance and coherence, hence their potential to effectively assist healthcare providers and patients.They did well in providing information on common conditions in cardiology and ENT, suggesting treatment options, and in frequently asked questions.The study also highlights the handling of ambiguous queries and conversation context conservation-prerequisites of the application of such systems in real-world healthcare settings.NLP integrated with chat and voice bot technologies enables support of patients, relieves workload pressure from healthcare professionals, and enhances the ultimate delivery of health services.In the future, this work will extend the dataset, then integrate more complex models of NLP like transformers, and finally, assure the privacy and security of data to lead to more accessible and efficient healthcare.
Language: English
Subfield: Health Informatics
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
OpenAlex ID: W4413441631