Machine learning models for reproducibility - initial release

Machine learning models for reproducibility

This repository contains trained machine learning models generated as part of different scientific studies. The goal is to support reproducibility, transparency, and external validation of published research by making the final models openly accessible.

Each study has its own folder containing one or more trained models. Models are GDPR-compliant and do not include raw, sensitive or personal data.

Included models:

• TED - TM use in cardiology: this model predicts telemedicine use (TM use) based on 41 predictor variables included in the TED study.

• Spartakus II - total illness duration until first diagnosis: this model predicts the total illness duration before the first diagnosis based on 183 predictor variables included in the Spartakus II study.

• Spartakus II - patient journey: this model predicts the patient journey class (poor/problematic, okay, or good/excellent) based on 183 predictor variables included in the Spartakus II study.

Documentation

See the README.md and model_info.md for:

• General information
• Model information
• Instructions to reuse the models

Release date

25 November 2025

Version

v1.0 – Initial stable version

Future updates will include additional machine learning models.