DIABOT - AI Powered Diabetes Chatbot

Abstract

Diabetes mellitus is a critical global health issue requiring continuous education, monitoring and support. In recent years, AI – driven technologies have transformed how healthcare is delivered, offering innovative solutions for disease management. This paper introduces Diabot, an intelligent conversational AI chatbot tailored for diabetes education and risk assessment. By combining Natural Language Processing (NLP), Machine Learning(ML), and Deep learning(DL). Diabot provides dynamic interactions, personalized advice, and early risk prediction. Integrated with message platforms like Telegram, it ensures accessibility and instant support for users, making it a promising tool in Al-powered preventive healthcare.

Introduction

Diabetes mellitus is a widespread chronic disease requiring ongoing management through medical care, education, lifestyle changes, and monitoring. Healthcare systems face challenges in providing timely and personalized support due to resource constraints. Advances in Artificial Intelligence (AI), particularly in Natural Language Processing (NLP) and Machine Learning (ML), have enabled the development of chatbots that offer scalable, interactive healthcare assistance.

This paper introduces Diabot, an AI-powered chatbot designed to help individuals manage diabetes by providing 24/7 support, symptom tracking, risk prediction, and personalized educational recommendations. Accessible via Telegram, Diabot combines conversational AI with predictive ML models to enhance diabetes self-care and awareness, especially in underserved areas.

The literature review highlights AI chatbots' growing role in healthcare, noting that while general health chatbots exist, few offer condition-specific diabetes support that integrates real-time messaging and personalized education. Existing diabetes tools mostly require manual data entry and lack conversational engagement.

Diabot’s development involved collecting and preprocessing health data from the Pima Indian Diabetes dataset and training multiple ML models, with XGBoost showing the best accuracy (85.2%) for diabetes risk prediction. NLP models like GPT-4 and BERT enable natural, context-aware user interactions. The system is deployed on Telegram using a Python Flask backend, integrating ML and NLP components.

User testing with 30 participants showed high satisfaction (4.5/5) for ease of use, response accuracy, clarity, and engagement. Diabot effectively provides diabetes-related advice and risk assessments. The chatbot’s accessible platform, instant responses, and friendly interface promote user trust and frequent interaction.

Limitations include inability to provide clinical diagnoses, dependence on internet connectivity, and challenges in handling complex medical questions. Future improvements aim to add multilingual support, device integration, and expanded medical knowledge.

Conclusion

This paper presented Diabot, an AI-Powered chatbot designed to support diabetes education, awareness, and early risk prediction. By Integrating machine learning algorithms with natural language processing and deploying the solution on a user-friendly platform like Telegram, Diabot offers a practical, scalable, and accessible digital health companion. The chatbot successfully provides: • Real Time, personalized responses to diabetes-related queries • Accurate risk assessment using trained ML models like XGboost • Seamless, natural interactions powered by advanced NLP techniques Experimental results demonstrate that Diabot is not only accurate in prediction but also well-received by users in terms of usability and effectiveness. It brings the gap between AL and healthcare, offering meaningful support to individuals who may lack direct access to medical professionals. In the future, Diabot can be enhanced by: • Supporting multiple languages • Connecting with wearable health devices (e.g., glucose monitors) • Offering integration with cloud-based electronic health records (EHRs) • Improving medical query handling using larger LLMs and custom-trained datasets Diabot serves a promising initiative in the field of AI-driven preventive healthcare, empowering users with knowledge, guidance and motivation for healthier living.

References

[1] Ali, A. 2001.Macroeconomic variables as common pervasive risk factors and the empirical content of the Arbitrage Pricing Theory. Journal of Empirical finance, 5(3): 221–240. [2] Basu, S. 1997. The Investment Performance of Common Stocks in Relation to their Price to Earnings Ratio: A Test of the Efficient Markets Hypothesis. Journal of Finance, 33(3): 663-682. [3] Bhatti, U. and Hanif. M. 2010. Validity of Capital Assets Pricing Model.Evidence from KSE-Pakistan.European Journal of Economics, Finance and Administrative Science, 3 (20).

Copyright

Copyright © 2025 Mr. S. Arun, Mr. Sanchi Sunil Kumar, Mr. Palakuru Lokabiram, Mr. R. Nitheesh, Mr. R. Uday Kiran, Dr. R. Karunia Krishna Priya, Mr. A. Venkatesan. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.