【作者】 柏林；

【导师】 陈竞先；

【作者基本信息】 东华大学 ， 管理科学与工程， 2006， 硕士

【摘要】 企业信息系统已在大多数企业中普及,企业数据也已经成为企业维持正常运营、积累竞争优势的一大资源。然而日益变化的企业内外部环境注定了企业对于数据的需求是一个不断增长的长期过程,而企业内部各组织、流程和地域之间的复杂关联,也导致企业对于数据的需求往往以分散和片面的形式呈现;同时,企业信息化的阶段性也加速了数据孤岛的一再出现,以上这些因素使得企业数据环境越来越趋向复杂化。 一个不良的企业数据环境造成的问题是显而易见的:一方面导致企业流程效率下降,信息系统的可信性降低,另一方面为数据仓库和数据挖掘等商务智能的应用带来了较大的风险和成本。目前在类似领域的国内外研究方法主要包括数据质量和数据集成两方面,在数据质量研究中,侧重数据质量的维度定义和度量手段等理论性方法,忽视了对数据适用性的探讨;在数据集成领域侧重对现存异构数据从语义和语法上寻求协调统一,从静态和局部上提供了多种解决方案。本文旨在结合企业实际存在的问题,提出企业数据合理更多还原

【Abstract】 Enterprise data is now becoming a important resource for business operation and corporation competitiveness development. As a result of changes in business environment, enterprise data requirement evolvement is a long term, ever-lasting progress. Besides, the existence of diverse data requirement between different department, business process and location leads to a distributed data storage and usage. And the fact that corporate information system is largely planned and implemented by phase also brings heterogeneity to data systems. These three factors makes enterprise data environment more and more sophisticated.A Bad-Organized data environment brings out many problems, including decreasing proficiency of business process, reducing the incredibility of information systems and adding cost and risk in data warehousing and data mining projects.Current approaches can be classified into two categories: data quality and data integration. Data quality approaches focus on quality dimension definition, measurement and improvement, with less method discussing on "Fitness for Use" of enterprise data;and approaches proposed in data integration literatures aim at solutions for data integrity both semantically and syntactically in heterogonous autonomic and distributed data. These approaches solve inconsistency and quality problem to some degree but often in static and partial way.This thesis proposed a composite approach Enterprise Data Rationalization Approach to fulfill enterprise data requirement and improve both data integrity and quality with dynamic and overall method. Enterprise data rationalization approach separates enterprise core data from operational data, adopting flexible data modeling techniques and central architecture for core data, and implement data enrichment workflow management together meta data repository and service to ensure core data’s quality.There are three data rationalization enables in the proposed approach: 1. Core data separation, central data repository and flexibledata modeling tehiniques altogether to meet themulti-dimensional and changing data requirement,2. Data enrichment workflow management system to ensure the quality and traceability of data process and provide workflow definition customization to align with process reengineering.3. Meta data repository and services to store and management business and technical knowledge and provide quality assurance.This thesis also gives an example implemented by a multinational company based on proposed approach, illustrating how to combine these three enablers to reengineer and rationalize corporation data environment.更多还原