BELB

BELB是一个专为生物医学实体链接设计的基准数据集，由柏林洪堡大学计算机科学系创建。该数据集包含11个语料库，涵盖基因、疾病、化学、物种、细胞系和变种等六种实体类型，并与7个专业知识库链接。BELB旨在通过统一格式减少预处理负担，提供标准化测试平台，以促进可重复性实验。数据集的应用领域包括生物医学文献的信息提取，旨在解决现有基准中缺乏此类标准化评估工具的问题。

BELB is a benchmark dataset specifically designed for biomedical entity linking, developed by the Department of Computer Science at Humboldt-Universität zu Berlin. This dataset includes 11 corpora covering six entity types: genes, diseases, chemicals, species, cell lines, and variants, and is linked to seven professional knowledge bases. BELB aims to reduce preprocessing burdens through a unified format and provide a standardized testing platform to facilitate reproducible experiments. Its application areas include information extraction from biomedical literature, and it is designed to address the lack of such standardized evaluation tools in existing benchmarks.