tanomanzo / mimic_attitude

Caregiver Attitude Detection from Clinical Notes

This project aims to detect the attitudes of caregivers from their clinical notes. This repository contains all the notebooks used for the study.

Installation

All notebooks work with Python 3.9. The required libraries are listed in the respective notebook, and can be installed using the following command:

pip install -r requirements.txt

Usage

Notebooks MIMIC_attitude_models_training_few_shot and MIMIC_attitude_models_training_few_shot contain the code for training and fine-tuning the models for few-shot and full-trained scenarios, respectively.

Notebooks zero_shot_mimic_sentiment and full_mimic_sentiment contain the code for testing the zero-shot and full-trained models, respectively. The few-shot scenario can be tested using a smaller percentage of the dataset used in full_mimic_sentiment.

The evaluation of the testing results can be found in zero_shot_mimic_sentiment_eval and full_mimic_sentiment_eval, respectively. The same evaluation can be conducted for the few-shot scenario.

The graphs used in the paper are generated in Graphs and ModelGraphs.

Finally, the annotation agreement can be found in AnnoAgreement.

The Hugging Face APIs for using the models can be found here: https://huggingface.co/tanoManzo/

License

The annotated data and the clinical notes follow the PhysioNet license. For information on the code token, please contact [email protected].