Table 2_Machine learning-based identification of inflammatory biomarkers for predicting pu

42 pediatric patient records from a retrospective case-control study conducted between January 2020 and December 2024. The data, published by Qianqian Dai, was used to identify three core inflammatory biomarkers for predicting pulmonary consolidation using five machine learning algorithms.

Use Cases

• Training a binary classifier to predict pulmonary consolidation based on lactate dehydrogenase (LDH), C-reactive protein (CRP), and erythrocyte sedimentation rate (ESR) levels.
• Benchmarking feature selection methods like LASSO, SVM-RFE, Random Forest, XGBoost, and LightGBM on a small clinical dataset.
• Developing a clinical risk assessment tool for early identification of high-risk pediatric patients with Chlamydia pneumoniae infection.
• Performing cluster analysis (e.g., K-means) to stratify patients into high-inflammation and low-inflammation groups based on key biomarkers.

Strengths

• Identifies three specific, consensus-validated predictive biomarkers (LDH, CRP, ESR) through five machine learning algorithms.
• Risk assessment system derived from the data demonstrated high predictive performance (AUC=0.993) and was validated with bootstrap resampling (1,000 iterations).
• Clear temporal coverage from January 2020 to December 2024.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is very small (23.6 KB), indicating a limited sample size of 42 patients.

Provenance

Source

figshare

Collection Method

Retrospective case-control study of 42 children with C. pneumoniae infection.

Time Range

January 2020 to December 2024

Freshness

Last updated 2026-05-04 05:33:15; freshness should be verified.

Geography

null