Code repository for Castela Forte et al. (2021) study on ML prediction models for prediction of 30-day, 1-year and 5-year mortality risk after cardiac surgery using pre-, intra-, and postoperative data
To run the project, install conda here. In order to install the conda environment, please run the following command in your terminal:
make setup-conda-environment
To run the model on a small peri-operative dataset with default parameter settings, use the following command:
make train_model_total
Otherwise, you can run from your terminal the model with some custom parameter settings. Use the following command to see those:
python predictor.py -h
You can also clean the model saves and plots:
make clean-results
- analyze_model_performance: folder of scripts for generating metric scores and plots
- data: data folder
- data500: data folder containing small dataset of 500 patients
- model: folder containing different models needed for training
- plots: folder containing different model plots
- preprocessing: folder containing preprocessing scripts
- stats: folder with model metric results
- training_data: folder containing model saves for each experiment