GACN

Please refer to the short paper we wrote regarding this work: Paper

Introduction

This project is a final project of NLP course.

Data

We took all the annotated data(train1, train2, test) combined it and randomly split to train and test.

** NOTE: The data for this project is not publicly available, for more details see: https://www.i2b2.org/NLP/Obesity/ **

Preprocessing, Training and Generating notes

first step:

preprocessing the data - run preprocessing.py

second step:

train and evaluate natural clinical notes classification - run classifier_train.py

third step:

train T5 model and generate new synthetic clinical notes - run generate_synthetic_clinical_notes.py

fourth step:

train and evaluate natural and synthetic clinical notes classification - run combined_calassifier_train.py