SmartHealthBot: A Sustainable AI Chatbot for Health Equity and Digital Innovation

Abstract

In many areas where there is little infrastructure available to deliver timely and reliable healthcare, users can remin an opportunity to use SmartHealthBot - an AI-powered healthcare chatbot providing real-time, individualized assistance, driven by artificial intelligence and deep learning. This utilizes natural language processing to interpret user symptoms provided via either voice or text format, allowing users\' interactions to be user friendly and inclusive. It leverages a hybrid methodology combining the advantages of both artificial neural networks and Naive Bayes classification, providing the ability to predict disease based upon a user\'s reported symptoms. Specifically, artificial neural networks (ANNs) are used to model the complex interrelationships between symptoms, while Naive Bayes classification provides for the rapid identifying of likely symptom groups. Further, through its use of sentiment analysis, it provides its users with mental health monitoring capabilities, including providing stress relief suggestions and motivation, as well as recommendations for enhancing diet and preventing illness, thereby supporting sustainable healthcare that encourages prevention. SmartHealthBot provides a rich analytics dashboard that assists the user in visualizing the trends and interactions of their symptoms, gaining greater insight into themselves.

Introduction

The text presents SmartHealthBot, an AI-powered healthcare chatbot designed to improve access to healthcare, particularly mental health support and early disease detection. Traditional healthcare often depends on face-to-face consultations, which can cause delays, while stigma and limited awareness prevent many people from seeking mental health care. AI, Machine Learning (ML), and Natural Language Processing (NLP) can help make healthcare more accessible, proactive, and personalised.

SmartHealthBot combines disease prediction, mental health monitoring, personalised lifestyle recommendations, diet planning, and health-record management in one platform. It uses NLP to support text and voice interaction and sentiment analysis to understand users' emotional states and provide motivational support.

Methodology

The system uses medical datasets containing diseases and associated symptoms. The data is cleaned, standardised, encoded into numerical values, and divided into 80% training and 20% testing data, with oversampling used to balance the dataset.

Two machine-learning models are developed:

• Artificial Neural Network (ANN): Used to identify complex relationships between symptoms and diseases.
• Naïve Bayes: Used for fast probability-based disease prediction.

The models are evaluated using accuracy, precision, recall, and F1-score, with the better-performing model intended for integration into the chatbot through NLP.

System Architecture

SmartHealthBot consists of several major components:

• Data preprocessing: Cleaning, standardisation, symptom encoding, tokenisation, lemmatisation, stop-word removal, TF-IDF vectorisation, normalisation, and dataset balancing.
• Disease prediction: ANN and Naïve Bayes models analyse symptoms and predict possible diseases.
• NLP: Enables users to communicate with the chatbot through text and speech.
• Recommendation system: Provides precautionary advice and personalised health recommendations.
• Mental health module: Tracks mood, analyses emotional trends, assesses risk levels, and suggests wellness activities.
• Health dashboard: Presents disease risks, symptom trends, mood, and nutritional information visually.

Dataset

The system uses two open-source datasets:

• Disease and Symptoms Dataset: Contains 4,920 records, 17 symptom attributes, and disease labels. Symptoms are converted into binary values for machine learning.
• Precautionary Dataset: Contains 41 diseases, with four precautionary measures for each disease, supporting the recommendation system.

Results

The ANN model achieved 96% accuracy, significantly outperforming the Naïve Bayes model, which achieved 78.12% accuracy. The ANN was therefore the stronger model for disease prediction.

The implemented system includes:

1. Login system for secure user access.
2. AI chatbot for symptom-based health assessment.
3. Health dashboard showing disease risks, symptom trends, mood, and nutrition.
4. Personalised diet planner based on medical conditions and dietary preferences.
5. Mental health module for mood tracking, emotion analysis, risk assessment, and wellness suggestions.

Conclusion

This project provides Intelligent Medical Assistance through an Interactive Chatbot that Analyses User Input Symptom Data and Predicts Possible Illnesses. This project is highly beneficial for patients because it uses Artificial Neural Networks (ANN), and the Naive Bayes method of Probabilistic Reasoning, giving users reliable predictive power to analyze Disease Symptoms and Recommendations. Because of this, SmartHealthBot uses AI algorithms for Deep Learning accuracy and Probabilistic Reasoning for Reliable Predictive Power, resulting in meaningful and useful recommendations for patients. SmartHealthBot combines Natural Language Processing (NLP) technology with user communications to provide more engaging conversations between patients and the healthcare system. In addition, SmartHealthBot has additional useful modules such as Diet Planning, Mental Health Support, and Health Records Management, which promote holistic health. SmartHealthBot provides a Personalised Healthcare Experience for Every User by generating Dietary Recommendations according to their Lifestyle Preferences, Evaluating Users\' Emotional States through Sentiment Analysis, and Saving Users\' Medical History.

References

[1] Ankit Jain, \"Mental Health Chatbot using Sentiment Analysis,\" IEEE Journal, 2023. [2] Prashant Biradar & Shivashankar Shastri, \"Medical Chatbot: AI-Based Infectious Disease Prediction Model,\" Journal of Scientific Research and Technology, 2024. [3] Sourav Chakraborty, \"A Review on NLP Applications in Medical Diagnosis,\" ACM Transactions on Computing for Healthcare, 2023. [4] Rahul Deshmukh, \"AI-Based Health Monitoring and Advisory System,\" Springer, 2024 [5] Shyam Dongre, Ritesh Chandra & Sonali Agarwal, \"MLtoGAI: Semantic Web Based with Machine Learning for Enhanced Disease Prediction and Personalized Recommendations using Generative AI,\" arXiv Preprint, 2024. [6] Aman Gupta, \"Healthcare Chatbots for COVID-19 Self-Assessment,\" IEEE Access Journal, 2021. [7] Pradeep Gorrepati, \"A Comparative Analysis of AI-Based Chatbots for Disease Diagnosis Based on Symptoms,\" International Journal of Intelligent Systems and Applications in Engineering, 2025. [8] Manish Jain, \"AI Healthcare Chatbot,\" Institute of Electrical and Electronics Engineers (IEEE), 2024. [9] Rohit Kumar, “AI-Driven Conversational Agents for Healthcare Support”, Journal of Biometric Informatics, 2023 [10] Cheng Liang, “MKA: A Scalable Medical Knowledge Assisted Mechanism for Generative Models on medical Conversation Tasks,” arXiv Preprint, 2023. [11] Ahmed Mohamed & Zhuopeng Li, “Artificial Intelligence Healthcare Chatbot System,” International Journal of Advanced Research in Computer and Communication Engineering (IJARCCCE), 2020. [12] Dhruv D. A. Parikh & Rakesh S. Nair, “A Review on Retinal Image Processing Techniques for Glaucoma Diagnosis,” Biometric Signal Processing and Control, 2021 [13] Rakesh Patel & Sanjay Sharma, “Personal Diet Recommendation using AI”, ACM Digital Library, 2022. [14] Karthik Prabhu, Anil Reddy & Tarun Raj, “Transforming Healthcare: An AI-Driven Medical Assistant Chatbot for Accurate Disease Diagnosis and Personalized Recommendations”, IJRASET, 2025 [15] Vijay Praveen & R Nagusundaram, “AI-Based Medical Chatbot for Disease Prediction,” International Journal of Science, Engineering and Technology (IJSET), 2024.

Copyright

Copyright © 2026 Mrs. V. Divya Raj, P. Keerthy, L. Amrutha, T. Nishitha, V. Meghana. 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.