Abstract

Diabetes mellitus is a chronic metabolic disorder affecting over 537 million adults globally, with prevalence expected to reach 783 million by 2045. Even with advances in diagnostic technology, equitable and early detection remains a challenge, particularly across demographically diverse populations. This study presents the design and development of an inclusive diabetes detection system that combines machine-learning classification, Local Interpretable Model-agnostic Explanations (LIME)-based explainability, and interactive Streamlit web deployment. A publicly available synthetic dataset of 100,000 patient records was preprocessed through encoding, Min-Max normalisation, and SMOTE-based class balancing. Four machine learning models (XGBoost, Random Forest, Naive Bayes, and Long Short-Term Memory (LSTM)) were trained and evaluated using accuracy, precision, recall, F1-score, and AUC-ROC. Hyperparameter tuning was conducted using Grid Search with 5-fold cross-validation. XGBoost achieved the best performance with 91.99% accuracy, precision of 1.0000, recall of 0.8665, F1-score of 0.9285, and AUC-ROC of 0.9438, clearly exceeding the research target of 70% accuracy. Demographic subgroup analysis across gender, ethnicity, age group, and income level confirmed equitable performance, with F1-score variation below 0.004 across most demographic dimensions. LIME generated clinically meaningful individual-level explanations, consistently identifying HbA1c and fasting glucose as the dominant predictive features. A broader evaluation was done, the trained XGBoost model was further tested on the Pima Indians Diabetes Dataset, a widely used benchmark with only 8 features, achieving 72.53% accuracy and AUC-ROC of 0.7969 using only 6 of 38 model features, demonstrating generalisation capability beyond the original training distribution. The complete system was deployed as a browser-based Streamlit application enabling real-time prediction and transparent explanation for clinical and non-clinical users.

References

• International Diabetes Federation. IDF Diabetes Atlas, 10th edition. Brussels, Belgium: IDF; 2021. Available at: https://diabetesatlas.org
• Maniruzzaman, M., Rahman, M.J., Ahammed, B. and Abedin, M.M., 2020. Classification and prediction of diabetes disease using machine learning paradigm. Health Information Science and Systems, 8(1), p.7.
• Rastogi, R. and Bansal, M., 2023. Diabetes prediction model using data mining techniques. Measurement: Sensors, 25, p.100605.
• Maniruzzaman, M., Rahman, M.J., Ahammed, B. and Abedin, M.M., 2020. Classification and prediction of diabetes disease using machine learning paradigm. Health Information Science and Systems, 8(1), p.7.
• Abnoosian, K., Farnoosh, R. and Behzadi, M.H., 2023. Prediction of diabetes disease using an ensemble of machine learning multi-classifier models. BMC Bioinformatics, 24(1), p.337.
• Hasan, M.K., Alam, M.A., Das, D., Hossain, E. and Hasan, M., 2020. Diabetes prediction using ensembling of different machine learning classifiers. IEEE Access, 8, pp.76516–76531.
• Zhou, H., Myrzashova, R. and Zheng, R., 2020. Diabetes prediction model based on an enhanced deep neural network. EURASIP Journal on Wireless Communications and Networking, 2020(1), p.148.
• Mahajan, P., Uddin, S., Hajati, F. and Moni, M.A., 2023. Ensemble learning for disease prediction: A review. Healthcare, 11(12), p.1808. MDPI.
• Cheungpasitporn, W., Athavale, A., Ghazi, L., Kashani, K.B., Colicchio, T., Koyner, J.L., Chen, J., Ix, J.H., Nadkarni, G. and Neyra, J.A., 2026. Transforming Nephrology Through Artificial Intelligence: A State-of-the-Art Roadmap for Clinical Integration. Clinical Kidney Journal, p.sfag004.
• Akther, F., Begum, M., Mahmud, T., Boltayev, A., Hanip, A. and Hossain, M.S., 2025. Streamlit-based AI for multi-disease prediction. In 2025 3rd International Conference on Inventive Computing and Informatics (ICICI), pp.1622–1628. IEEE.
• Febrian, M.E., Ferdinan, F.X., Sendani, G.P., Suryanigrum, K.M. and Yunanda, R., 2023. Diabetes prediction using supervised machine learning. Procedia Computer Science, 216, pp.21–30.
• Arumugam, K., Naved, M., Shinde, P.P., Leiva-Chauca, O., Huaman-Osorio, A. and Gonzales-Yanac, T., 2023. Multiple disease prediction using machine learning algorithms. Materials Today: Proceedings, 80, pp.3682–3685.
• Bukhari, M.M., Alkhamees, B.F., Hussain, S., Gumaei, A., Assiri, A. and Ullah, S.S., 2021. An improved artificial neural network model for effective diabetes prediction. Complexity, 2021(1), p.5525271.
• Ihnaini, B., Khan, M.A., Khan, T.A., Abbas, S., Daoud, M.S., Ahmad, M. and Khan, M.A., 2021. A smart healthcare recommendation system for multidisciplinary diabetes patients with data fusion based on deep ensemble learning. Computational Intelligence and Neuroscience, 2021(1), p.4243700.
• Krishnamoorthi, R., Joshi, S., Almarzouki, H.Z., Shukla, P.K., Rizwan, A., Kalpana, C. and Tiwari, B., 2022. A novel diabetes healthcare disease prediction framework using machine learning techniques. Journal of Healthcare Engineering, 2022(1), p.1684017.
• Singh, S., Agarwal, S., Singh, V., Hazela, B., Dubey, V. and Singh, P., 2025. Comparative analysis of machine learning algorithms for multiple disease prediction model with an optimized scalable deployment. In 2025 International Conference on Electronics, AI and Computing (EAIC), pp.1–6. IEEE.
• Beghriche, T., Djerioui, M., Brik, Y., Attallah, B. and Belhaouari, S.B., 2021. An efficient prediction system for diabetes disease based on deep neural network. Complexity, 2021(1), p.6053824.
• Sharma, A., Guleria, K. and Goyal, N., 2021. Prediction of diabetes disease using machine learning model. In International Conference on Communication, Computing and Electronics Systems: Proceedings of ICCCES 2020, pp.683–692. Singapore: Springer Singapore.