CRM系统中维度建模的应用研究

【作者】 陈兵飞；

【导师】 陈清华；

【作者基本信息】 南京理工大学 ， 计算机应用技术， 2006， 硕士

【摘要】 近年来，啤酒工业利用信息技术生产和搜集数据的能力大幅提高，大量的数据库被用于管理、办公、科学研究和工程开发。如何高效地利用信息资源，保留现有客户，开拓新市场，提高企业的核心竞争力，成为啤酒工业信息化建设的核心问题。本文主要通过对实施南通大富豪啤酒工业分析型CRM系统的经验总结，对基于数据仓库、面向啤酒工业的分析型CRM系统的设计和实现进行研究和总结。针对啤酒工业的特点，在维度设计中，采用了分裂为微型维度的方法设计了大型变化客户维，有效地解决了大型维度的历史信息记录和查询优化的问题；在分析模型设计中，提出了改进的五度客户分割法，科学地分割了客户，为实现对客户提供个性化服务奠定了基础。在OLAP数据存储上，设计和实现了MOLAP和ROLAP相结合的方法，兼顾了高效的数据存储和快速的查询响应的优点。同时，对数据挖掘技术在啤酒工业的应用进行了展望，并对顾客满意度测试模型进行了研究。更多还原

【Abstract】 Recently, as the ability that beer industry utilizes the information technology to product and collect data is improved by a wide margin, lots of databases are used in the management, official working, and scientific research and project development. As a result, How to utilize infonnation resources, keep existing customer, open up the new market effectively and improve the key competitiveness of enterprises, is becoming the core problem of industrial information construction of beer.This paper is, mainly through the summary of the experience which implements the industrial analytical CRM system of beer of nantong dafuhao, to research and summarize the design and implementation of the analytical CRM system of beer industry, which is based on data warehouse.In dimensional modeling, design the large-scale change customer dimension by splitting it into some mini dimensions degree. It has solved the problem of recording historical information and optimizing inquiry about large-scale change customer dimension effectively; in analysis modeling, propose a improved method of splitting customer into five degrees, cut apart the customer scientifically and establish the foundation for realizing to the personalized service. In OLAP data store, design and realize the method combined MOLAP with ROLAP. It has given consideration to the advantage that the high-efficient data store and fast inquiry respond. Meanwhile, look forward to the application of data mining technology in beer industry and research the testing modeling of contentment by customer.更多还原