A synthetic electronic health record dataset designed to contain realistic clinical signals. The dataset is described as enabling the training of models achieving a 0.97 ROC-AUC score. It originates from Kaggle.
Use Cases
- Benchmarking classification algorithms based on the described high ROC-AUC performance.
- Training models for clinical signal detection based on the presence of realistic signals.
- Testing synthetic data generation techniques for EHR privacy preservation.
- Developing healthcare predictive models using synthetic patient data.
Strengths
- Dataset is described as enabling models with a 0.97 ROC-AUC score.
- Contains realistic clinical signals according to the description.
Limitations
- Row count, column names, and file formats are unknown.
- Description metadata is limited; actual data quality requires manual inspection after download.
- Last update date is unknown; freshness unverified.