Body Composition Phenotypes in 78 Obese Children Aged 6–18 Years

A dataset of 78 obese children aged 6–18 years from a single-center outpatient clinic, used to identify body composition-based obesity phenotypes. The data includes 13 BIA-derived indices analyzed via unsupervised clustering and interpretable machine learning. It was authored by Yuhang Wang and last updated in June 2026.

Use Cases

• Identifying distinct body composition phenotypes based on 13 BIA-derived indices.
• Training random forest classifiers for phenotype prediction using features like fat mass and BMI.
• Applying SHAP analysis to interpret the contribution of features like lean mass and skeletal muscle mass to cluster separation.

Strengths

• Includes data from 78 participants, a specific cohort size for an exploratory study.
• Analysis achieved a cross-validated classification accuracy of 85.9% and a macro AUC of 0.953.
• Cluster stability was evaluated with a high adjusted Rand index of 0.94 and Jaccard indices ranging from 0.78 to 0.86.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data is from a single-center, cross-sectional design, which the authors note is exploratory and requires external validation.

Provenance

Source

figshare, author Yuhang Wang.

Collection Method

Cross-sectional study enrolling 78 obese children from a single-center outpatient clinic; body composition measured via Bioelectrical Impedance Analysis (BIA).

Freshness

Last updated 2026-06-03 11:11:49; freshness should be verified.