The notebook includes code to train four different models—a Random Forest, an XGBoost, an LSTM, and a CNN-LSTM model—to predict a patient’s future glucose levels based on historical glucose data. The dataset used in this project is the OhioT1DM dataset from Ohio State University, which provides detailed data on individuals with Type 1 diabetes. Additionally, the notebook includes commands to download the dataset.
The trained models for each patient are available for download at this link.