PheBee Performance Benchmarks: Synthetic Subject-Phenotype Datasets

Five synthetic clinical phenotype datasets (1,000 to 100,000 subjects) with clinically plausible disease clustering, term frequency, and qualifier distributions, designed for reproducible performance evaluation of the PheBee clinical phenotype knowledge graph system. Generated using Human Phenotype Ontology (HPO) terms.本数据集为五组合成临床表型数据集，样本量覆盖1000至100000名受试者，具备符合临床合理性的疾病聚类、词频与限定符分布特征，旨在为PheBee临床表型知识图谱系统的可复现性能评估提供支撑。该数据集基于人类表型本体（Human Phenotype Ontology, HPO）术语生成。