NMR Diabetes Prediction

The repository containing the code for the thesis titled: "Machine Learning Based Survival Analysis Integrating Blood NMR Data for Explainable Type 2 Diabetes Prediction".

Setting up environment

To get started, please install the requirements on a Python 3.10 environment. An example using conda:

conda create -n venv python=3.11
conda activate venv
pip install -r requirements.txt

Structure

The repository is structured as follows:

• results_fs contains all results regarding the feature selection experiment (Experiment 1).
• results_models contains all results regarding the model experiments (Experiments 2-4).
• methods contains all plots used in the thesis methods.
• src contains all code used to run all experiments.

Usage notice

Due to the proprietary nature of the datasets used in the thesis, none were included here. For the sake of reproducability and transparency, it was decided to still publish the code. If any code or analysis is found to be incorrect, please feel free to open an issue or contact the author.