Patient-specific blood glucose prediction using deep learning, considering the challenges of "small dataset" and "data imbalance"

The official implementation of the paper "Gluformer: Transformer-Based Personalized Glucose Forecasting with Uncertainty Quantification."

Source code for predicting Blood Glucose Concentration

The official implementation of the paper "GlucoBench: Curated List of Continuous Glucose Monitoring Datasets with Prediction Benchmarks."

GluPredKit aims to make blood glucose model training and prediction more accessible.

"Personalised Short-Term Glucose Prediction via Recurrent Self-Attention Network" in CBMS 2021

Official code for PhysioCGM: A Multimodal Physiological Dataset for Non-Invasive Blood Glucose Estimation (Nature Scientific Data)

Benchmark of glucose predictive models in diabetes

Implementation of architecture for 2020 OhioT1D competition submission. Includes weights from pre-training runs with Tidepool data set. Baseline architecture is N-BEATS, modifications include RNN/shared output blocks, additional Losses. https://folk.idi.ntnu.no/kerstinb/kdh/KDH_ECAI_2020_Proceedings.pdf

Converter from different popular CGM data export formats and datasets to a unified format suitable for ML training and inference

Thie is the repository for preprocessing the D1NAMO dataset

"Jointly Predicting Postprandial Hypoglycemia and Hyperglycemia Using Continuous Glucose Monitoring Data in Type 1 Diabetes" in EMBC 2023

Free and open-source Dexcom desktop app with notifications, trends, and Nightscout support

Data Science Project - Predicting glucose levels with data collected by non-invasive wearable device

Deep Learning for Continuous Glucose Monitoring Prediction — Temporal Fusion Transformer, Clarke Error Grid analysis, hypoglycemia prevention for Abbott Libre/Medtronic MiniMed

Machine learning web app to predict user glucose levels

A time-series forecasting extension for pydexcom using Google's TimesFM

Previsione a breve termine della glicemia nel diabete di tipo 1 con modelli di machine learning interpretabili.

End-to-end blood-glucose forecasting platform: 3-phase tiered ML (CGM / post-CGM / watch-only Virtual CGM) with rigorous offline evaluation (baselines, clinical metrics, conformal intervals, LOSO, drift), a FastAPI + CockroachDB backend, a React Native app, and a keyless local-LLM health assistant. Portfolio project, APK-only.