Blood Glucose Prediction - Research Platform

[miriamkw]

Tasks remaining to finish the tool:

We will build heavily on the benchmarkpaper for this tool, to test it on real life scenarios.

Preparations

• Create a research branch, from which you will use as base branch for this work
• Remove the irrelevant code (models, metrics, documentation, example files)

Parsers
Summary: All the parsers should do the same: return a "raw dataframe" with all of the datatypes in the same time grid, but without any fuzz.

• Rewrite tidepool parser to do this
• Rewrite nightscout parser to do this
• Create an apple health parser since it has heartrate as input
• Create an Oura Ring parser?
• On call, save the dataframe in the data/raw repository (title metadata: start and end date, parser used, username?)
• Start making the CLI: Add an option to parse data
• Document: Nighscout cannot provide workouts, tidepool can, document which dataypes from apple parser

Preprocessors

• Define all prediction horizons --> add target columns
• Define how much history to use
• ALWAYS IN MG/DL!
• Save the df with a very descriptive name
• Test, train and val split
• Test: Does it work with both NS and tidepool?
• Update documentation (README, CLI comments)
• Add preprocessors for more (all?) models in benchmarking paper
• Add what-if events in the preprocessor(s)?
• Right now, preprocessors assums numerical features to be cgm, insulin and carbs. This should be dynamic from user input (like in model training).

Models

• Implement some sample prediction models. Base the implementation on the benchmark paper
• Add as input the numerical and the categorical features
• Make sure that all the examples are working
• Add CLI and have proper commenting
• Add models from benchmarking paper:
  • scikit models
    • ARX (LinearRegressor)
    • Elastic net
    • Huber
    • Lasso
    • Random forest
    • Ridge
    • SVR, linear kernel
    • SVR, radial basis kernel
  • XGBoost
    • Gradient boosting trees
  • Keras
    • LSTM
    • TCN
• Add documentation in readme

Evaluation

• Make sure they are all implemented, conventional ones.
  • ESOD
  • TG
• Handle comparisons and single ones. Store results in files!
• Plots: Handle comparisons

Real-time plots:

• Adding plots with solutions to predict one specific trajectory, either in real time or for a given date (with "true" measurements alongside).
• Adding interactive plot?

Settings

• Use global settings for unit conversion
• Implement all relevant files to have this as a boolean input. Fetching from config in the main script.

Command line interface

• Main menu: Settings or model
• Choose either preprocessing or existing dataset
• Choose either model training or pretrained model
• Choose evaluation metrics

Documentation

• Make it really clear in the beginning what this repository is, and shortly explain how it works
• Make it clear how to start the program
• Make it clear how to get the raw data: either by having your own dataset on a specific format, or
• Document datatypes
• Document the flow of the predictions
• Document the CLI
• Document file structure and base models. Make clear how people can add their:
  • models
  • evaluation metrics
  • input values without using parsers
  • pretrained models
• Document evaluation metrics
• Document plots
• Document assumptions:
  • 5-minute intervals of CGM inputs

Code improvements (feedback from chatgpt):

• Add default values to config-file
• Dynamic choices for click-options (listing alternatives from modules)
• Check for potential exceptions (input verification), and write clear error messages
• Output verbose information: adding --verbose flag to describe cli commands
• add helper methods in a separate file to the CLI to keep the CLI file clean
• Add unit testing
• Consistency: When using models/metrics etc, make sure you use the same type of CLI-argument
• Duplicate code: Find a way to fix that. It is in processors, model training ...