WALVIS

WALVIS是一个针对荷兰语的生物医学实体链接数据集，由莱顿大学和伊拉斯姆斯大学医学中心的研究人员自动生成。该数据集通过整合UMLS、Wikidata和Wikipedia的内容，无需专家手动标注，实现了对生物医学实体的自动链接。数据集包含2400个独特的实体提及，这些提及均与UMLS中的概念相对应。WALVIS数据集主要用于训练和评估荷兰语的生物医学实体链接模型，特别是在处理患者生成的文本时，能够进行高级别的分析。WALVIS is a Dutch-language biomedical entity linking dataset automatically generated by researchers from Leiden University and Erasmus University Medical Center. It integrates content from UMLS, Wikidata and Wikipedia, enabling automatic biomedical entity linking without manual expert annotation. The dataset contains 2,400 unique entity mentions, all of which correspond to concepts in UMLS. The WALVIS dataset is primarily used for training and evaluating Dutch biomedical entity linking models, and supports high-level analysis especially when processing patient-generated text.