BioREx

BioREx是由美国国家生物技术信息中心开发的一个大型生物医学关系抽取数据集，旨在通过整合多个异构数据集来提高关系抽取的性能。该数据集包含600篇PubMed文章的标注，涵盖了基因、疾病、化学物质等多种生物医学概念之间的关系。BioREx的创建过程涉及对不同数据集的统一格式化和特征调整，以确保数据的一致性和可用性。该数据集主要应用于生物医学文献中的知识发现和知识图谱构建，旨在解决生物医学文本处理中的关系抽取问题。

BioREx is a large-scale biomedical relation extraction dataset developed by the National Center for Biotechnology Information (NCBI), which aims to enhance the performance of relation extraction by integrating multiple heterogeneous datasets. This dataset includes annotated data from 600 PubMed articles, covering relations between diverse biomedical concepts such as genes, diseases, and chemical substances. The development of BioREx involves uniform formatting and feature adjustment across disparate datasets to ensure data consistency and usability. This dataset is primarily applied to knowledge discovery and knowledge graph construction in biomedical literature, targeting the resolution of relation extraction issues in biomedical text processing.