bigbio/msh_wsd
收藏资源简介:
MSH WSD数据集是一个用于生物医学领域词义消歧(WSD)评估的数据集。该数据集通过使用统一医学语言系统(UMLS)和MEDLINE的手动MeSH索引自动生成。数据集包含106个模糊缩写、88个模糊术语和9个两者的组合,总计203个模糊词。每个实例都分配了UMLS 2009AB版本的CUI,每个模糊词/缩写的每个意义最多包含100个实例,总计37,888个模糊案例,来自37,090个MEDLINE引用。
The MSH WSD dataset is a benchmark dataset for word sense disambiguation (WSD) evaluation in the biomedical domain. It is automatically constructed using the Unified Medical Language System (UMLS) and manually annotated MeSH indices from MEDLINE. The dataset encompasses 106 ambiguous abbreviations, 88 ambiguous terms, and 9 combinations thereof, amounting to a total of 203 ambiguous lexical items. Each instance is assigned a Concept Unique Identifier (CUI) from the UMLS 2009AB release. Each sense of every ambiguous term or abbreviation contains up to 100 instances, resulting in an overall total of 37,888 ambiguous cases derived from 37,090 MEDLINE citations.
数据集概述:MSH WSD
基本信息
- 语言: 英语
- 许可证: UMLS_LICENSE
- 多语言性: 单语种
- PubMed可用性: 是
- 公开性: 否
- 任务类型: 命名实体消歧(NAMED_ENTITY_DISAMBIGUATION)
数据集详情
- 主页: https://lhncbc.nlm.nih.gov/ii/areas/WSD/collaboration.html
- 数据组成: 包含106个模糊缩写、88个模糊术语和9个混合类型,总计203个模糊词汇。
- 数据来源: 使用UMLS Metathesaurus和MEDLINE的手动MeSH索引自动生成。
- 数据量: 共37,888个模糊案例,分布在37,090篇MEDLINE文献中。
引用信息
@article{jimeno2011exploiting, title={Exploiting MeSH indexing in MEDLINE to generate a data set for word sense disambiguation}, author={Jimeno-Yepes, Antonio J and McInnes, Bridget T and Aronson, Alan R}, journal={BMC bioinformatics}, volume={12}, number={1}, pages={1--14}, year={2011}, publisher={BioMed Central} }




