MicrobioRel

该数据集是一个特定领域的语料库，专为微生物组相关的生物医学文献中的多类别关系提取而设计。它包含了来自科学文章的手动标注摘录，覆盖了六大实体类别中的22种关系类型。此外，该数据集还包括了一个“None”类别，以支持负例的学习，并且所有命名实体都预先使用PubTator进行了标注，并由领域专家进行了验证。该数据集的规模为1,994个标注关系，其任务是进行多类别关系提取。

This dataset is a domain-specific corpus tailored for multi-class relation extraction from microbiome-related biomedical literature. It contains manually annotated excerpts from scientific articles, covering 22 relation types across six entity categories. Additionally, the dataset includes a "None" category to support negative example learning, and all named entities were pre-annotated using PubTator and validated by domain experts. The dataset consists of 1,994 annotated relations, with its dedicated task being multi-class relation extraction.