Development of a Time-Series Deep Learning Framework for Real-Time Prediction of Hypoglycemia and Hyperglycemia

Mohammad Shafeeq; Monika Tripathi

doi:10.55524/ijircst.2026.14.4.5

Abstract

Predicting blood glucose variability accurately is essential in managing type 1 diabetes mellitus (T1DM). The occurrences of hypoglycemic and hyperglycemic events can be managed much better if detected early on, therefore preventing severe adverse events. Often existing deep learning methods either cannot capture local glucose variation patterns or do not model long-term temporal dependencies from the continuous glucose monitoring (CGM) data. To overcome this limitation, this study presents a novel multi-scale convolutional neural network-bidirectional long short-term memory-transformer (MSCNN-BiLSTM-Transformer) hybrid framework for multi-class glucose state prediction. In sequential glucose measures? the proposed model aims at learning a discriminative representation through multi-scale convolution-based feature extraction, bidirectional temporal learning and transformer based self-attention mechanism. The OhioT1DM dataset was utilized in experiments that employed 12 previous glucose readings as inputs to classify glucose states into three states, hypoglycemia, normal and hyperglycemia. The proposed model achieve an accuracy of 97.59%, precision of 97.78%, recall of 97.56%, F1 score of 97.60%, ROC-AUC of 99.79%, Matthews correlation coefficient (MCC) of 95.30%, Cohen’s Kappa 95.76%, and outperform other conventional CNN, LSTM, GRU, Transformer and other baseline models. This work presents an ablation study that showed our multi-scale convolution, bidirectional sequence learning and Transformer contextual modelling and feature fusion is effective. Moreover, the SHAP-based explainability analysis sheds light on how historical glucose data affected the final prediction, demonstrating the proposed framework's clinical reliability and interpretability. The results obtained from the tests show that the revealed hybrid model is a robust, accurate, and interpretable model for intelligent glucose state prediction which has the potential for being used as a decision-support tool for personalized diabetes treatment and other smart healthcare solutions in the future.

Keywords

Type 1 Diabetes Mellitus, Continuous Glucose Monitoring, Blood Glucose Prediction, Multi-Scale CNN, BiLSTM, Transformer, Explainable Artificial Intelligence, SHAP, Deep Learning, Healthcare Analytics

References

O. Vershinina et al., “Explainable artificial intelligence model predicting the risk of all-cause mortality in patients with type 2 diabetes mellitus,” Frontiers in Endocrinology, vol. 16, Art. no. 1689312, Oct. 2025. Available from: https://doi.org/10.3389/fendo.2025.1689312
H. Chen and L. Zhao, “Federated learning for privacy-preserving diabetes prediction in smart healthcare systems,” Artificial Intelligence in Medicine, vol. 147, Art. no. 102611, 2025. Available from: https://doi.org/10.1016/j.artmed.2025.102611
N. Patel and S. Verma, “Explainable AI for diabetes prediction using SHAP with ensemble models,” Journal of Biomedical Informatics, vol. 146, Art. no. 104466, 2024. Available from: https://doi.org/10.1016/j.jbi.2024.104466
R. Kumar, M. Singh, and A. Sharma, “IoT-enabled diabetes monitoring system using LSTM and wearable sensors,” IEEE Internet of Things Journal, vol. 10, no. 1, pp. 785–793, 2023. Available from: https://doi.org/10.1109/JIOT.2023.3250148
Y. Zhang, X. Li, and Y. Wang, “Hybrid deep learning for blood glucose prediction using continuous glucose monitoring data,” Biomedical Signal Processing and Control, vol. 86, Art. no. 105286, 2023. Available from: https://doi.org/10.1016/j.bspc.2023.105286
M. Garcia, J. Perez, and R. Alvarez, “Transformer-based blood glucose forecasting for diabetes management,” Computer Methods and Programs in Biomedicine, vol. 244, Art. no. 107945, 2024. Available from: https://doi.org/10.1016/j.cmpb.2024.107945
X. Liu, H. Wang, and Y. Chen, “Deep temporal neural networks for personalized glucose prediction using CGM data,” IEEE Journal of Biomedical and Health Informatics, vol. 28, no. 4, pp. 2195–2206, 2024. Available from: https://doi.org/10.1109/JBHI.2024.3365127
A. Singh, P. Kumar, and R. Gupta, “Multi-horizon blood glucose prediction using attention-enhanced BiLSTM networks,” Knowledge-Based Systems, vol. 302, Art. no. 112867, 2025. Available from: https://doi.org/10.1016/j.knosys.2025.112867
T. Zhou, H. Li, and L. Chen, “Explainable transformer networks for continuous glucose forecasting,” Expert Systems with Applications, vol. 270, Art. no. 126134, 2025. Available from: https://doi.org/10.1016/j.eswa.2025.126134
J. Wang, Z. Huang, and Y. Sun, “Multivariate time-series prediction of blood glucose using hybrid CNN-LSTM models,” Applied Soft Computing, vol. 140, Art. no. 110215, 2023. Available from: https://doi.org/10.1016/j.asoc.2023.110215
S. Ahmed, M. Islam, and M. Hasan, “Deep learning-based prediction of hypoglycemia events using wearable sensor data,” Sensors, vol. 24, no. 3, Art. no. 1134, 2024. Available from: https://doi.org/10.3390/s24031134
J. Kim, S. Lee, and H. Park, “Personalized glucose prediction using transformer and attention mechanisms,” IEEE Access, vol. 12, pp. 53142–53156, 2024. Available from: https://doi.org/10.1109/ACCESS.2024.3389211
F. Rodrigues, P. Silva, and M. Costa, “Real-time prediction of glucose abnormalities using sequential deep learning models,” Scientific Reports, vol. 15, Art. no. 12487, 2025. Available from: https://doi.org/10.1038/s41598-025-12487-9
T. Prioleau, B. Lu, and Y. Cui, “Glucose-ML: A collection of longitudinal diabetes datasets for development of robust AI solutions,” Scientific Data, vol. 12, Art. no. 987, 2025.
B. Cinar, L. van den Boom, and M. Maleshkova, “Deep learning-based hypoglycemia classification across multiple prediction horizons,” arXiv, 2025. Available from: https://arxiv.org/abs/2504.00009
B. Cinar, L. van den Boom, and M. Maleshkova, “A review on machine learning approaches for the prediction of glucose levels and hypoglycemia,” arXiv, 2026. Available from: https://arxiv.org/abs/2601.11615
G. Rigamonti, M. P. Barbato, D. Marelli, and P. Napoletano, “Tailoring adverse event prediction in type 1 diabetes with patient-specific deep learning models,” arXiv, 2026. Available from: https://arxiv.org/abs/2601.14917
L. Obiuwevwi, K. J. Rechowicz, V. Ashok, and S. Jayarathna, “Towards affordable, non-invasive real-time hypoglycemia detection using wearable sensor signals,” arXiv, 2026. Available from: https://arxiv.org/abs/2602.10407
B. Cheng et al., “A 65 nm trustworthy hypoglycemia forecasting engine achieving 11.3 nJ per inference,” arXiv, 2026. Available from: https://arxiv.org/abs/2606.07455
Y. Luo, F. Chen, and W. Zhang, “Hybrid transformer–LSTM framework for continuous glucose forecasting in diabetes management,” Neural Networks, vol. 183, Art. no. 106935, 2025. Available from: https://doi.org/10.1016/j.neunet.2025.106935
A. H. Zuberi et al., “Efficacy of structured exercise protocol on functional capacity, HbA1c, and waist-hip ratio in post-myocardial infarction subjects with diabetes mellitus,” Journal of Neonatal Surgery, vol. 14, no. 5S, pp. 396–405, 2025. Available from: https://doi.org/10.52783/jns.v14.2058
S. K. Singh, P. Verma, and P. Kumar, “Sentiment analysis using machine learning techniques on Twitter: A critical review,” Advances in Mathematics: Scientific Journal, vol. 9, no. 9, pp. 7085–7092, 2020. Available from: https://doi.org/10.37418/amsj.9.9.58
S. K. Singh, P. Verma, and P. Kumar, “Sentiment analysis of COVID-19 epidemic using machine learning algorithm on Twitter,” Journal of Critical Reviews, vol. 7, no. 18, pp. 2565–2572, 2020. Available from: https://www.jcreview.com/
S. Joshi, P. Gupta, R. Kumar, A. Dwivedi, and S. Kumar Singh, “Privacy by Design: User-Centric Control in Online Photo Sharing,” in 2025 International Conference on Intelligent and Secure Engineering Solutions (CISES), Greater Noida, India, 2025, pp. 1287–1290. Available from: https://doi.org/10.1109/CISES66934.2025.11265161
M. Mishra, M. S. Hussain, and S. K. Singh, “Protecting Against Social Engineering Using Wireshark: Effective Strategies With Real-World Examples,” in Effective Strategies for Combatting Social Engineering in Cybersecurity. Hershey, PA, USA: IGI Global, 2025. Available from: https://doi.org/10.4018/979-8-3693-6665-3.CH008
G. Mishra, H. Dev, and S. K. Singh, “Comparative Analysis of Pre-Trained CNN and Feature Selection Algorithms for Plant Leaf Disease Identification,” in 2024 IEEE 11th Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON), Lucknow, India, 2024, pp. 1–7. Available from: https://doi.org/10.1109/UPCON62832.2024.10983102
G. Mishra, H. Dev, and S. K. Singh, “Evolutionary Algorithms and Deep Learning Techniques for Plant Disease Identification: A Comprehensive Review,” in Recent Advancements in Semiconductor Technologies, Lecture Notes in Electrical Engineering, vol. 1491. Singapore: Springer, 2026. Available from: https://doi.org/10.1007/978-981-95-3305-3_27
A. Singh, H. Agarwal, A. P. Singh, A. Kumar, S. Srivastava, and S. K. Singh, “Deep learning and red deer optimiser for automatic cardiovascular disease identification on magnetic resonance images,” Scientific Reports, 2026. Available from: https://doi.org/10.1038/s41598-026-55595-6

Cites this article as

M. Shafeeq, M. Tripathi, "Development of a Time-Series Deep Learning Framework for Real-Time Prediction of Hypoglycemia and Hyperglycemia", International Journal of Innovative Research in Computer Science and Technology (IJIRCST), Vol-14, Issue-4, Page No-44-60, 2026. Available from: https://doi.org/10.55524/ijircst.2026.14.4.5

Corresponding Author

Mohammad Shafeeq

Department of Computer Science and Engineering, P K University, Shivpuri, Madhya Pradesh, India

Download Full Paper

Download PDF

No. of Downloads: 4 | No. of Views: 30

Development of a Time-Series Deep Learning Framework for Real-Time Prediction of Hypoglycemia and Hyperglycemia

Citations

Total View 30

Total Download 4