This repository contains code that attempts to re-create the results of Tan et al in "4SDrug: Symptom-based Set-to-set Small and Safe Drug Recommendation". The code contained within attempts to achieve the same results as the aformentioned paper with special care taken to avoid plagiarizing the code base created for the previous work.
The "Overleaf project" will house all written work.
- torch
- dill
- numpy
- scipy
Run the main.py file to run our code.
Data is sourced from the MIMIC website. However, we use the preprocessed data found in the original code base linked above.
Section 3 of the paper details the requirements of the model. We use the original implementation.