Repository for paper on evaluation metrics for privacy-preserving synthetic data. We utilized Python 3.11.

• install required packages: pip install -r requirements.txt
• create a data/ directory
• download the following tables from MIMIC_IV v3.1 as gzipped csv and put them in the data/ directory: admissions, patients, omr, diagnoses_icd
• run utils/preprocess_mimic.py to acquire the dataset.
• run main.py (alter number of cross validation folds, evaluation metrics, etc. as seen fit)