MeSHup

MeSHup是由西安大略大学和多伦多大学联合创建的大型生物医学文档索引数据集，包含1,342,667篇英文全文文章及其相关的MeSH标签和元数据。该数据集从MEDLINE数据库中收集，旨在通过提供丰富的全文信息，支持生物医学文本分类和信息检索的研究。MeSHup数据集的创建过程涉及从PubMed Central Open Access和MEDLINE/PubMed年度基准库中下载和匹配文章，确保数据的质量和多样性。该数据集广泛应用于自动支持生物医学文献索引的研究，旨在解决手动索引过程中的时间和成本问题。MeSHup is a large-scale biomedical document indexing dataset jointly developed by the University of Western Ontario and the University of Toronto. It contains 1,342,667 full-text English articles along with their associated MeSH terms and metadata. Collected from the MEDLINE database, this dataset aims to support research on biomedical text classification and information retrieval by providing rich full-text information. The creation process of the MeSHup dataset involves downloading and matching articles from PubMed Central Open Access and the annual baseline repository of MEDLINE/PubMed to ensure the quality and diversity of the dataset. This dataset is widely used in research on automatically supporting biomedical literature indexing, aiming to address the time and cost issues in manual indexing processes.

搜集汇总

背景与挑战

背景概述

MeSHup是一个大规模标注的生物医学文档索引语料库，包含来自MEDLINE数据库的1,342,667篇全文文章，每篇文章都带有MeSH标签和元数据（如作者和出版场所）。该数据集旨在支持自动化MeSH索引系统的开发和评估，解决手动标注效率低下的问题，并提供了完整的训练和使用指南。