RobotPsychologist / bg_control

Improving short-term prandial blood glucose outcomes for people with type 1 diabetes, a complex disease that affects nearly 10 million people worldwide. We aim to leverage semi-supervised learning to identify unlabelled meals in time-series blood glucose data, develop meal-scoring functions, and explore causal machine-learning techniques.
https://blood-glucose-control.streamlit.app/
18 stars 42 forks source link

First pass at PyODAnnotator model. #95

Closed gavinkatz001 closed 1 week ago

gavinkatz001 commented 4 weeks ago

Seeking criticism on this first attempt at the PyODAnnotator model. I know I am doing something incorrectly based on my results, just not sure what.