摘要
[目的/意义]随着大数据环境下医疗信息化的飞速发展,医学数据类型和规模也不断增加。面对医学信息在该过程中出现的冗余、异构等现象,通过词表间映射进行知识组织系统的互操作可以实现语义消歧和概念逻辑上的统一。[方法/过程]以《中国中医药学主题词表》TC类、《中国图书馆分类法》R类向《中文医学主题词表》语义映射为例,依靠深度学习工具Word2Vec为技术手段,实现了实验对象词条的向量形式转化。在此基础上根据词向量相似度结果与目标词表类目进行自动化匹配筛选,建立映射。[结果/结论]基于Word2Vec进行的映射能够在一定程度上实现互操作,其思路可为在类似的知识组织系统间建立语义关联时提供参考,在精确性和方法的综合运用上仍存在着提升空间。
[Purpose/significance]With the rapid development of medical informatization under the big data environment,the type and scale of medical data and information is also increasing.In the face of redundancy and heterogeneity of medical information in the process,semantic disambiguation and conceptual logic unification can be achieved by interoperability of knowledge organization systems through word list mapping.[Method/process]Taking Semantic Mapping from Traditional Chinese Medicine Thesaurus(TC)and CLC(R)to CMeSH as an example,relying on the deep learning tool Word2 Vec technology,realizing the transformation of experimental object vector form.On the basis of the word vector similarity result and target word category automatic matching mapping.[Result/conclusion]Mapping based on Word2 Vec can realize the interoperation,the idea can provide a reference for establishing semantic associations between similar knowledge organization systems,and more improvement needs to be made in accuracy and comprehensive application of methods.
出处
《情报理论与实践》
CSSCI
北大核心
2019年第9期160-165,176,共7页
Information Studies:Theory & Application
基金
国家自然科学基金重大项目“国家安全大数据综合信息集成与分析方法”的研究成果之一,项目编号:71790612
