Pickled XGBoost models + Jupyter notebook (code)

This repository contains the code and model supporting the article "Machine-learning based prediction of appendicitis for patients presenting with acute abdominal pain at the emergency department". The notebook with the code has had all sensitive information removed.

The model, an XGBoost, was developed in Python 3.8, using the XGBoost version 2.0.3 using the DMLC package (https://xgboost.readthedocs.io/en/stable/python/index.html).

The model requires input parameters from: intake information, vital signs, medical history and physical examination. Please consult the paper and supplementals for exact input parameters.

The model has been developed using patient data from one Dutch teaching hospital (2016-2023) using 336 cases with 80/20 percent training-test split. This model is a proof-of-concept, and it is strongly advised to develop it further using more data to ensure robustness and generalizability to other clinical contexts and patient populations.