COMPARATIVE STUDY OF INDEX TERMS AND FULL-TEXT

越来越多的生物医学期刊正在制作中 可通过全文进行在线搜索，即完整的 对每篇文章的文本进行排序，并且可以通过以下方式检索任何文章 包含在其中的单个单词或单词的组合。这 表示信息访问点的急剧增加 在这些日记里。 但是，在某些情况下，全文可用性可能会显示 接入点过多，阻碍了MOST的识别 相关文件。另一方面，受控指数项， 代表对下列重要概念的判断 文件和判决都会受到人为错误的影响。 该项目旨在测试指数的相对有效性 用于检索其中的文档的术语和全文 Medline期刊，也可以进行全文搜索。 将向以下部门征求信息请求 医学、外科、家庭和社区医学以及儿童 密苏里大学哥伦比亚健康科学分校的健康 中心。训练有素和经验丰富的搜索者将搜索 Medline数据库和相应的全文数据库 相关引文。将对检索到的所有引文进行评估 请求者的关联性。结果将以表格形式显示，并 分析以确定哪种检索方法更多 有效地识别相关的书目引用，为什么完整- 文本系统缺少一些相关文档，以及是否有 方法对于某些类型的 问题。此外，还将分析数据以确定如何 通常，全文系统中的相关条目可能 仅与标题和抽象词联系在一起。

An increasing number of biomedical journals are being made available for online searching via full-text, that is, the complete text of each article is sorted and any article may be retrieved by a single word or combinations of words contained therein. This represents a dramatic increase in points of access to information in these journals. In some cases, however, full-text availability may present a surfeit of access points and hinder the identification of the most relevant documents. Controlled index terms, on the other hand, represent judgments regarding the important concepts in documents, and judgments are subject to human error. This project is designed to test the relative efficacy of index terms and full-text for the retrieval of documents in those MEDLINE journals for which full-text searching is also available. Information requests will be solicited from the Departments of Medicine, Surgery, Family and Community Medicine, and Child Health at the University of Missouri-Columbia Health Sciences Center. Trained and experienced searchers will search both the MEDLINE database and the corresponding full-text databases for relevant citations. All citations retrieved will be evaluated for relevance by the requesters. Results will be tabulated and analyzed to determine which of the retrieval methods is more effective in identifying relevant bibliographic citations, why full- text systems miss some relevant documents, and whether one method is more effective than the other for certain kinds of questions. In addition, data will be analyzed to determine how often relevant items in full-text systems could have been identified with title and abstract words alone.