electricsheepafrica/african-neonatal-sepsis-dataset

该数据集为非洲新生儿败血症风险分层提供合成的数据集，使用来自49项研究的元分析中的基于证据的优势比生成。数据集包含验证的风险因素，如APGAR评分、低出生体重、早产和胎膜早破。这些数据集用于出生风险评分、早期发病与晚期发病分类以及死亡率预测。数据集旨在实现高AUC-ROC值，以预测败血症和死亡率，适用于资源匮乏环境中的分诊算法。

This dataset provides synthetic datasets for neonatal sepsis risk stratification in African populations, generated using evidence-based odds ratios from a meta-analysis of 49 studies involving 87,548 neonates. The datasets include validated risk factors such as APGAR score, low birth weight, preterm birth, and prolonged rupture of membranes. These datasets enable birth risk scoring, early-onset vs late-onset classification, and mortality prediction. The datasets are expected to achieve high AUC-ROC values for sepsis prediction and mortality risk, making them suitable for triage algorithms in under-resourced settings.