/BD4H-project-code-submission

Primary LanguagePython

Steps to run

  • ./preprocessing_caml/my_dataproc.ipynb : run this to create raw data that combines discharge notes and medical encoding label in the same row, and create train, validation, test split
  • ./Explainable-Automated-Medical-Coding-master: main code for the model in embeddings/ folder: run the code in playground.ipynb to create Word2Vec model for word and code embedding in HLAN/HAN_train.py : main training script. We use VSCode to launch training, see ./vscode/launch.json for hyperparameters and settings

Data

Since data is quite large, we are not submitting data folders with this repo The post processing data should be put in /Explainable-Automated-Medical-Coding-master/datasets/data The Word2Vec embedding model should be put in //Explainable-Automated-Medical-Coding-master/embeddings