This is a project to predict glucose by learning data collected by wearable devices and food logs.
With collected data(Accelerometer, Blood volume pulse, Electrodermal activity, Temperature, Interbeat interval, Heart rate, Food Log, Interstitial glucose concentration), feature engineering is performed to utilize meaningful features for learning.
Led this project as a data science lecture team project.
From: 24.05.20
To: 24.06.09
Period: 2024.07.01 ~ 2024.08.23 (5th Week ~ 8th Week)
Final Presentation Google Slide
- exploring_digital_biomarkers_a_day.ipynb : Data Exploration(for a day) Phase Related Code
- feature_engineering.ipynb : Code of the Data Preprocessing & Feature Engineering (Feature Extraction) step
- xgboost_regressor.ipynb : XGBoost Regressor(Cross Validation Methods: Leave One Subject Out, Partial Personalization)
- SHapley Additive ex-Planations (SHAP) results: force_plot | heatmap | summary_plot per a day
- Prediction plots
Cho, P., Kim, J., Bent, B., & Dunn, J. (2023). BIG IDEAs Lab Glycemic Variability and Wearable Device Data (version 1.1.2). PhysioNet. https://doi.org/10.13026/zthx-5212.
Bent, B., Cho, P.J., Henriquez, M. et al. Engineering digital biomarkers of interstitial glucose from noninvasive smartwatches. npj Digit. Med. 4, 89 (2021). https://doi.org/10.1038/s41746-021-00465-w