%0 Journal Article
%T Web数据转换模式映射优化方法<br/>Web Data Exchange Schema Mapping Optimization Method
%A 纪宇航
%A 李贵
%A 李征宇
%A 韩子扬
%A 曹科研
%J Hans Journal of Data Mining
%P 76-89
%@ 2163-1468
%D 2020
%I Hans Publishing
%R 10.12677/HJDM.2020.101008
%X Web数据转换是Web异构数据源集成的重要研究之一，通常分为实例层和模式层两方面进行。本文的研究主要针对模式层，由于给定的源到目标模式映射通常使数据转换结果包含大量冗余，为了生成不含冗余的数据作为数据转换核解，本文设计了一种基于同态关系的模式映射设计与优化方法。该方法首先引入模式映射之间的同态关系作为模式映射重写方法基础，通过对模式映射进行分解，定义不同规则生成的数据冗余的大小程度，确定需要重写的规则。最后将给定的模式映射重写为能够直接生成核解的核模式映射，并将其转换为可执行的SQL语句来计算核解。本文实验使用来自中国土地市场网的数据验证本文方法的有效性。<br />
Web data exchange is one of the important researches on the integration of Web heterogeneous data sources. It is usually divided into two aspects: instance layer and schema layer. The research in this paper is mainly focused on the mode layer. Because a given source-to-target mode mapping usually makes the data exchange results contain a lot of redundancy, in order to generate data without redundancy as a data exchange kernel solution, this paper designs a homomorphic rela-tionship Schema mapping design and optimization methods. This method first introduces the ho-momorphic relationship between the schema mappings as the basis of the schema mapping re-writing method. By decomposing the schema mappings, defining the degree of data redundancy generated by different rules, and determining the rules that need to be rewritten. Finally, the given schema mapping is rewritten into a kernel schema mapping that can directly generate a kernel so-lution, and it is converted into an executable SQL statement to calculate the kernel solution. This paper uses data from China Land Market Network to test the performance of the proposed method.
%K Web大数据，数据转换，模式映射，核解，同态关系<br/>Web Big Data
%K Data Exchange
%K Schema Mapping
%K Core Solution
%K Homomorphism
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=33936