Skip to content

Latest commit

 

History

History
44 lines (30 loc) · 2.39 KB

README.md

File metadata and controls

44 lines (30 loc) · 2.39 KB

Machine Learning & Pediatric Appendicitis

This repository holds the code for the analysis of the data from pediatric patients with suspected appendictis. Machine learning (ML) models are trained and validated to perform the prediction of the diagnosis, management, and severity. The resulting models were deployed as a research prototype of the Pediatric Appendicitis Prediction Tool. Further details can be found in the Frontiers in Pediatrics paper.

This reserach was carried out by the Medical Data Science research group, ETH Zurich, Switzerland, in collaboration with the University Children's Hospital Regensburg (KUNO), Germany, Research and Development Campus Regensburg (WECARE), Germany, and Department of Pediatric Surgery and Pediatric Orthopedics, Hospital St. Hedwig of the Order of St. John, Regensburg, Germany.

Requirements

To run this code, you will need

  • R, version ≥ 3.6.2
  • RStudio, version ≥ 1.1.456

All the necessary packages can be installed by running setup_script.R.

Data

The data are provided in CSV and RDA fromats in app_data.csv and app_data_clean.Rda files, respectively.

Usage

  • setup_script.R sets up necessary packages and utility functions
  • eda.R performs some basic exploratory analysis on the dataset
  • diagnosis.R, management.R, and severity.R validate logistic regression, random forest, and gradient boosting models for predicting diagnosis, management, and severity, respectively
  • variable_selection.R performs variable selection based on the random forest variable importance

Maintainer

This repository is maintained by Ričards Marcinkevičs ([email protected]).

Citation

If you use the dataset, please cite the paper:

@article{MarcinkevicsWolfertstetter2021,
  title={Using Machine Learning to Predict the Diagnosis, Management and Severity of Pediatric Appendicitis},
  author={Marcinkevics, Ricards and Reis Wolfertstetter, Patricia and Wellmann, Sven and Knorr, Christian and Vogt, Julia E},
  journal={Frontiers in Pediatrics},
  volume={9},
  pages={360},
  year={2021},
  publisher={Frontiers}
}