Published Paper Details:

In developing countries, the issue of accessibility to healthcare is serious because of language diversity, the lack of medical workers, disorders in the emergency system, and the high cost of treatment. This study presents a Multilingual Artificial Intelligence-Based Healthcare Support System that is meant to improve the initial disease detection process, enhance healthcare information, organize emergency response, and facilitate telemedicine. The system has a multilingual Natural Language Processing (NLP) system that enables users to communicate in English, Hindi, and Marathi both through text and voice. Its disease prediction model uses an ensemble machine learning model based on a combination of the Random Forest, Support Vector machine, and Deep Neural Network models, and is trained on a dataset consisting of 773 diseases, 377 symptoms. The system was implemented with a microservice-based architecture (Python and FastAPI) and containerized with Docker and Kubernetes, which guarantees its scalability, reliability, and high availability. According to experimental results, the ensemble model can predict with an accuracy of 88.2 percent and a high F1-score, in addition to system-level testing showing that the ensemble model can operate steadily even with more than 100,000 simultaneous users, response time of less than 400 ms, and multilingual processing accuracy of over 92 percent. This suggested platform is a scalable and viable solution to deliver early medical advice, emergency management, and remote medical care, especially in resource-limited settings.