Ensemble-learning approach improves fracture prediction using genomic and phenotypic data
Authors' conclusion
The ensemble learning approach significantly improved fracture prediction accuracy and model performance compared to individual ML.
Study context
- Population
- Older men in the Osteoporotic Fractures in Men cohort StudyMrOS cohort
- Participants
- 5,130 men
- Comparison
- Super Learner (SL) ensemble model vs individual machine learning (ML) algorithms and baseline models
- Outcome
- Prediction of major osteoporotic fractures (MOF), evaluated on the testing set
Methodological detailsensemble machine learning (Super Learner), random split into training (80%) and testing (20%) sets, tenfold cross-validation, subgroup analyses by race
- Ensemble machine learning (Super Learner)
- Random split into training (80%) and testing (20%) sets
- Tenfold cross-validation
- Subgroup analyses by race
Performance
As reported in the abstract
MOF prediction performance (testing set)
| Test or model | AUC | Accuracy | Sensitivity | Specificity | NPV | PPV |
|---|---|---|---|---|---|---|
| SL | 0.76 | 95.6% | 94.5% | 96.1% | 95.1% | 94.7% |
SL
AUC0.76
Accuracy95.6%
Sensitivity94.5%
Specificity96.1%
NPV95.1%
PPV94.7%
Further findings
| Outcome | SL |
|---|---|
| Accuracy by race (subgroup) | |
| Accuracy, Whites | 93.1% |
| Accuracy, Minorities | 91.6% |
Accuracy by race (subgroup)
- SL outperformed single ML in subgroup analysis
Original publication
Ensemble-learning approach improves fracture prediction using genomic and phenotypic data. Osteoporos Int. Published 7 March 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →