Data Sheet 1_Lipid metabolism-based machine learning models for predicting large for gesta

A dataset of 1,968 singleton pregnancies without diabetes, used to develop machine learning models for predicting large for gestational age (LGA). The data includes lipid profiles measured at early (11–14 weeks) and mid-pregnancy (20–24 weeks). The dataset was created by Wanqing Liu and is available under a CC-BY-4.0 license.

Use Cases

• Training machine learning models for LGA risk prediction based on pre-pregnancy BMI and gestational weight gain.
• Evaluating the predictive capacity of lipid metabolism parameters, such as mid-pregnancy triglycerides.
• Conducting feature importance analysis using SHAP values to identify key clinical predictors.
• Comparing the performance of logistic regression, random forest, SVM, and XGBoost algorithms on clinical data.

Strengths

• Contains data from 1,968 patient records, split into training (n=1,574) and independent test (n=394) sets.
• Models were developed using nested 5x5 cross-validation to prevent overfitting, with stable convergence reported.
• The best-performing XGBoost model achieved an AUC of 0.776 on the independent test set.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the full dataset is unknown, which may limit suitability assessment.
• The dataset is small in scale at 350.9 KB, indicating limited scope.

Provenance

Source

figshare

Collection Method

Retrospective analysis of singleton pregnancies without diabetes.

Freshness

Last updated 2026-05-15 04:26:27; freshness should be verified.