Blood Glucose Level Prediction in Type 1 Diabetes Using Machine Learning

TL;DR

This paper tackles short-term blood glucose prediction for Type 1 Diabetes by leveraging the DiaTrend CGM dataset to forecast BG at a 30-minute horizon using 15 methods spanning traditional ML, deep learning, deep reinforcement learning, and ensemble techniques. It provides a comprehensive comparison across these paradigms and glycemic states, highlighting that a voting ensemble combining MLP, LSTM, and GRU (V2) delivers the lowest RMSE overall, while LSTM and stacking excel in normoglycemia and DDPG leads in hypoglycemia. The study also employs a $NRMSE$-based reward for DRL and reports detailed performance across normo-, hyper-, and hypoglycemic conditions, demonstrating meaningful differences in method suitability by context. The findings have implications for refining closed-loop diabetes management and CGM-driven insulin dosing, and they emphasize the value of ensemble approaches in enabling robust BG predictions across diverse glycemic scenarios.

Abstract

Type 1 Diabetes is a chronic autoimmune condition in which the immune system attacks and destroys insulin-producing beta cells in the pancreas, resulting in little to no insulin production. Insulin helps glucose in your blood enter your muscle, fat, and liver cells so they can use it for energy or store it for later use. If insulin is insufficient, it causes sugar to build up in the blood and leads to serious health problems. People with Type 1 Diabetes need synthetic insulin every day. In diabetes management, continuous glucose monitoring is an important feature that provides near real-time blood glucose data. It is useful in deciding the synthetic insulin dose. In this research work, we used machine learning tools, deep neural networks, deep reinforcement learning, and voting and stacking regressors to predict blood glucose levels at 30-min time intervals using the latest DiaTrend dataset. Predicting blood glucose levels is useful in better diabetes management systems. The trained models were compared using several evaluation metrics. Our evaluation results demonstrate the performance of various models across different glycemic conditions for blood glucose prediction. The source codes of this work can be found in: https://github.com/soon-jynn-chu/t1d_bg_prediction

Figures (7)

• Figure 1: Distribution for Each Condition for Train and Test Set
• Figure 2: DNN Model Architecture Illustration
• Figure 3: ML methods prediction visualization
• Figure 4: NN methods prediction visualization
• Figure 5: RL methods prediction visualization