数据挖掘技术在电信IP决策支持系统中的应用

【作者】 袁新程；

【导师】 陆建峰；

【作者基本信息】 南京理工大学 ， 模式识别与智能系统， 2004， 硕士

【摘要】 随着电信业改革的深入和入世的影响，国内电信企业面临着日益激烈的竞争，在这样的环境下，能否快速做出科学的决策，对企业的发展有着重大的影响。 本文针对电信IP业务种类多，更新快的特点，设计了一个面向IP业务的综合决策支持系统。本系统使用SQL Server 2000构造了面向电信IP业务的星型数据仓库；构建了基于．NET框架的多层架构的系统框架，系统在具体实现时对各层又适当地进行模块化，并用．NET组件和COM／COM+组件实现，提高系统的重用性和扩展性；本系统用关联规则分析和聚类分析等数据挖掘算法对电信计费系统和经营系统的历史数据进行分析，并针对电信IP业务数据的特点对算法进行改进，最后将挖掘出的知识嵌入到系统中为相关决策提供帮助。更多还原

【Abstract】 Further reform in telecom business, the impact of joining the WTO, and the following demanding competition among domestic telecom enterprises are leading a modern telecom enterprise to the efficient and fast decision-making, which would be essential for the future development of the enterprise itself.This paper describes an IP business-oriented summing-up decision-making supportive system, which is designed and based on its various and fast updating characters in IP business. This system uses SQL Server 2000 to construct a star schema Data Warehouse (DW), which is facing the telecom IP business. It also builds the multi-layer system framework based on . NET framework. And the system modularizes every layer properly while specializing with .NET Model and COM/COM+ to enhance the expansibility of the system. This system adopts association analysis and clustering analysis algorithm to analyses the historical data of the telecom fee system and management system, improves the original algorithm according to the characters of data in telecom IP business, and finally offers help for the relative decision-making by embedding the gained knowledge into the system.更多还原