基于语义分析的元数据管理工具研究

发布时间：2018-06-23 11:08

本文选题：数据中心 + 元数据　；参考：《复旦大学》2012年硕士论文

【摘要】：目前,在数据中心建设中,来自各业务系统的海量数据没有一致和统一的数据定义,需要元数据管理工具提供元数据的统一视图,对这些数据来源、运作和变迁等情况进行跟踪分析。传统的元数据管理工具缺乏通用性和规范性,并且在语义分析方面也有先天的缺陷,因而研究基于语义分析的通用、规范,元数据管理工具具有重要意义。分析和设计了基于公共仓库元模型(Common Warehouse Metamodel,CWM)的元数据管理工具的元数据管理流程；在对工具的元模型管理、元数据管理和语义分析三方面进行分析的基础上,设计了工具的体系结构。所设计的工具主要包含四个部分：元模型管理模块、元数据管理模块、语义分析模块和元数据存储模块。借鉴Web本体语言(Web Ontology Language,OWL)语义表达和推理领域的成果,设计元数据信息向OWL的转化规则并证明了其正确性,制定了语义推理规则格式和设计语义推理引擎。在实现技术方面,元模型管理模块支持图形界面建模、模型批量导入导出和增量更新等功能；对于元数据管理模块,给出元数据建立和关系维护算法；对于语义分析模块,在Jena基础上实现的语义推理引擎,支持自定义语义推理规则,并利用其对元模型进行语义一致性检查等功能,给出元数据的血统分析算法、影响分析算法和差异分析算法。
[Abstract]:At present, in the construction of the data center, there is no consistent and uniform data definition for the massive data from various business systems, and metadata management tools are required to provide a unified view of the metadata. Follow up and analyze the operation and changes. The traditional metadata management tools are lack of generality and standardization, and also have inherent defects in semantic analysis. Therefore, it is of great significance to study the general, standardized and metadata management tools based on semantic analysis. The metadata management process of metadata management tools based on Common Warehouse MetamodelCWM (Common Warehouse Metamodel CWM) is analyzed and designed, and the metadata management, metadata management and semantic analysis are analyzed. The architecture of the tool is designed. The tool includes four parts: metamodel management module, metadata management module, semantic analysis module and metadata storage module. Based on the results of semantic representation and reasoning in Web Ontology language (owl), the transformation rules from metadata to owl are designed and proved to be correct. The format of semantic reasoning rules and the design semantic reasoning engine are developed. In terms of the implementation technology, the meta-model management module supports the functions of graphical interface modeling, model batch import and export and incremental update; for metadata management module, metadata establishment and relational maintenance algorithm; for semantic analysis module, The semantic reasoning engine based on Jena supports self-defined semantic reasoning rules, and uses the functions of semantic consistency checking of metamodel to give the algorithms of metadata lineage analysis, influence analysis and difference analysis.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP308;TP391.1

【参考文献】