XGBoost-SHAP-based interpretable diagnostic framework for knee osteoarthritis: a population-based retrospective cohort study
Authors' conclusion
Our study highlights the usefulness of machine learning in unveiling important factors that influence the diagnosis of KOA to guide new prevention strategies.
Study context
- Population
- Participants of the Wu Chuan knee osteoarthritis (KOA) Study, Inner Mongolia, China; anonymized questionnaire data26.3% diagnosed with KOA
- Participants
- 1,188 participants
- Exposure
- Four machine learning (ML) classifiers compared; best model explained with Shapley values
- Outcome
- Classification performance of ML models for diagnosing knee OA (sensitivity, specificity, PPV, NPV, accuracy, AUC, G-means, F1 scores)
Methodological detailspopulation-based, machine learning model development, 7:3 training/test split, cross-validation, class balancing
- Population-based
- Machine learning model development
- 7:3 training/test split
- Cross-validation
- Class balancing
Main finding
As reported in the abstract
Classification performance of the best model
| Group | AUC | G-means | F1 score |
|---|---|---|---|
| XGBoost with Boruta | 0.758 | 0.800 | 0.703 |
XGBoost with Boruta
AUC0.758
G-means0.800
F1 score0.703
Observational data: associations, not effects.
Further findings
Feature importance (SHAP)
- Top features of KOA by importance ranking
- 17
- Most important feature: average of the experience of joint pain
Original publication
XGBoost-SHAP-based interpretable diagnostic framework for knee osteoarthritis: a population-based retrospective cohort study. Arthritis Res Ther. Published 19 December 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →