基于关键字的数据元语义描述方法  

Semantic description method of data elements based on keywords

作　　者：胡青宁 董金平[2] 李婷玉 田源 苏宏伟 HU Qingning;DONG Jinping;LI Tingyu;TIAN Yuan;SU Hongwei(The Exploration Design and Information Center Institue,PetroChina Jidong Oilfield Company,Tangshan,Hebei 063004,China;Technology and Information Dept,PetroChina Jidong Oilfield Company,Tangshan,Hebei 063004,China)

机构地区：[1]中国石油冀东油田分公司勘察设计与信息化研究院,河北唐山063004 [2]中国石油冀东油田分公司科技信息处,河北唐山063004

出　　处：《东北石油大学学报》2022年第3期107-114,I0008,共9页Journal of Northeast Petroleum University

基　　金：中国石油冀东油田分公司科技攻关项目(JDYT-2020-JS-50311)。

摘　　要：针对油田区域数据湖跨业务域数据资源集成时,需要解决数据项之间的同名异义、异名同义问题,提出基于关键字的数据元语义描述方法。利用数据元中的对象词、特性词、表示词、限定词作为关键字,采用关键字集合表达数据元的语义,描述数据模型的语义,提出基于数据元名称编辑距离相似度与关键字项集相似度相结合的数据元语义相似度计算方法,将相关方法应用到油田区域数据湖建设中,给出数据元字典构建方法和基于关键字的数据元语义的数据模型映射方法。结果表明:基于关键字的数据元语义描述方法,不需要领域本体,简单易用,可提高数据映射的效率,适合工程化应用。Integrating the data resource across business domains in oilfield regional data lake,it is necessary to solve the problems of homonymy and synonymy between data items.This paper proposes a method of data element semantic description based on keywords,which takes object words,characteristic words,expression words and qualifiers as keywords,and uses keyword set to express the semantics of data element.A data element semantic similarity calculation method is proposed with combination of data element name edit distance similarity and keyword item set similarity.These methods are applied to the construction of oilfield regional data lake with the method of the data element dictionary construction and the data model mapping method based on the semantic of data element keywords.Practice shows that the semantic description method of data element based on keywords is simple and easy to use,which does not need auxiliary information such as domain ontology,can improve the efficiency of data mapping,and is suitable for engineering application.

关 键 词：数据元 关键字 语义树 语义相似度 数据映射 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]