【作者】 孙燕；

【导师】 刘巍；

【作者基本信息】 大连海事大学 ， 应用数学， 2004， 硕士

【摘要】 随着计算机技术、网络技术、通讯技术和Internet技术的发展，企业业务操作流程日益自动化，营销过程中产生了的大量的客户数据，这些数据和由此产生的信息是企业的宝贵财富，但是面对如此海量的客户数据，管理者却面临数据丰富而知识贫乏的困境。在客户关系管理中迫切需要寻找一种新的工具，来对企业的营销规律进行研究和探索，为营销决策提供有价值的知识，使企业获得更大的利润，其中数据挖掘便是一种很有发展前景的技术。 数据挖掘是从存放在数据库、数据仓库或其它信息库中的大量数据中挖掘有趣知识的过程，主要是基于人工智能、机器学习、统计学等技术，高度自动化的分析企业的数据，做出归纳性的推理，从中挖掘出潜在的模式，预测客户的行为，帮助决策者调整市场策略，做出正确的决策。粗糙集理论是一种新的有效的数据挖掘的方法。 属性约简是粗糙集用于数据分析的重要概念。寻求一种高效的属性约简算法仍然是粗糙集的研究热点之一，现尚不存在一种非常有效的约简算法。研究表明，最小约简的计算和全部约简的计算都是NP问题。在人工智能中，解决这类问题的一般方法是利用启发式信息进行约简。本文详细的分析了几种典型的约简算法，提出了一种新的基于物元的属性约简算法。从算法的执行时间、约简的完备性和约简的最小性几方面进行对比分析，通过实例说明了该方法的有效性和优越性。同时举例将基于物元的属性约简算法应用于客户关系管理的客户价值分析中得到合理有指导性的规则，并将此方法与可拓集合、物元变换、物元的可拓性相结合应用于新客户发掘中，给出了一种客户分类和新客户发掘的方法。更多还原

【Abstract】 Along with the development of computer network communication and Internet , the operation of the enterprise is becoming more automatic and then vast datum of the clients come into being . But in face of the vast datum the manager is facing the mess of indigent knowledge .In CRM it is imperious to find a new method to study and search after the rule in distribution of the commodity and offer valuable knowledge to the manager’s decision-making , and the enterprise will win more profit. Data Ming is a useful method .Data Ming is the process of discovering interesting knowledge from large amounts of Data stored either in Databases , Data Warehouses , or other information repositories . It is mostly based on artificial intelligence machine study and statistics . Data Mining is used to analyze the information of the clients automatically and make out a inductive consequence . Then we can find the latent mode to forecast the actions of the clients and help the manager make a correct decision . Rough Set is a new useful method to Data Ming .Attribute Reduction is a important concept used in data analysis of Rough Set . Finding a fast and useful Attribute Reduction arithmetic is one of the hot topics . It is said that the calculation of the smallest reduction and all reductions is a NP problem . In this paper we analyze a few kinds of Attribute Reduction arithmetic in detail and give a new arithmetic based on matter element . Then these kinds of arithmetic are compared from the carrying out time and maturity and briefness . By an example we have proved that the arithmetic based on matter element is effective and predominant. In an example , the arithmetic is used to analyze the value of the clients in CRM to find reasonable and directive rules , and we combine the method with Extension Set , matter element commutation and. matter element’s extension character to find new latent clients , offer a method of classing the clients and finding the latent clients .更多还原