Obesity-Depression Comorbidity Risk Factors from Korean and US Health Surveys

Data from 3,357 physically inactive adults in the Korea National Health and Nutrition Examination Survey (KNHANES, 2007–2012) was used to develop a predictive model. The XGBoost model was validated internally with KNHANES 2011–2012 data and externally with the U.S. NHANES dataset. The dataset, authored by Yuwen Shangguan, is associated with a study published on figshare.

Use Cases

• Training predictive models for obesity-depression comorbidity based on physical inactivity and metabolic features.
• Conducting feature importance analysis for health risk factors based on SHAP values mentioned in the description.
• Validating machine learning models across different national health survey datasets.
• Investigating the association between metabolic markers like insulin concentration and mental-physical health comorbidities.

Strengths

• Includes data from 3,357 individuals, providing a substantive sample for model development.
• Model performance is quantified with internal validation AUCs of 0.783 and 0.744 and an external validation AUC of 0.886.
• The dataset is associated with a study that performed external validation on a separate U.S. dataset, suggesting methodological rigor.
• Released under a permissive CC-BY-4.0 license, facilitating reuse.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the underlying data is unknown, which may limit suitability assessment.
• The 1.3 MB file size suggests a small dataset, potentially limiting model complexity or analysis scope.

Provenance

Source

Korea National Health and Nutrition Examination Survey (KNHANES) and U.S. NHANES.

Collection Method

Data were obtained from physically inactive adults enrolled in KNHANES (2007–2012).

Time Range

2007–2012 for primary data, with validation on 2011–2012 data.

Freshness

Last updated 2026-05-28 11:38:37; freshness should be verified.

Geography

South Korea and the United States.