Implementation of BernGraph: Probabilistic Graph Neural Networks for EHR-based Medication Recommendations
- Please check the requirements.txt
- You can create the conda environment and use the requirements.txt file:
pip install -r requirements.txt
- data.csv: preprocessed raw data
- label.csv: Ground truth of recommendation
- run.py: train BernGraph
- egsage.py: GNN model
- eval.ipynb: Evaluation file.
- data.csv: preprocessed raw data
- label.csv: Ground truth of recommendation
- run.py: train BernGraph
- egsage.py: GNN model
- eval.ipynb: Evaluation file.
First, please unzip the data files inside each folder.
- Training: Run the "run.py" file in each folder.
- Model parameter: The best parameter will be saved in the folder "/models".
- Evaluation: Run the "eval.ipynb" file in each folder.