数据挖掘在物流企业客户分析中的研究与应用.docVIP

摘 要 近几年来，第三方物流企业，竞争越来越激烈各个企业要想在激烈的竞争中处于优势地位，自己的客户。数据挖掘技术应用到企业，找出有价值的数据，从而维持老客户，新客户，提高客户的满意度，进而增加客户忠诚度以提高客户的价值和利润就成了一件竞争的利器，对加强物流企业的竞争力会起到十分重要的作用。3)该算法在预测时必须知道从叶子节点到根节点的路径上所有内节点对应属性的属性值。 针对缺点1)，本文用粗糙集理论中的属性依赖度对ID3算法进行改进，提出了RID3算法，RID3算法是依据决策属性对条件属性的依赖度k的大小，作为建立决策树的属性选择标准。针对缺点3),本文提出在用ID3算法构造决策树的同时，记录下从每个分类节点到根节点的路径上的分类数目，也就是说记录下每个分类节点处所包含的记录数。这样，即使测试集中的记录中存在缺失属性，我们也可以根据决策树来预测该事件发生的概率，从而提高了决策树的预测能力。在以上两点思想的基础上，本文提出了RID3-1算法，并用具体实例证明了该算法的有效性，以及相对于ID3算法的优越性。 （2）本文通过对某综合型物流企业的客户数据和业务情况进行了分析，确立了分析的主题，在Oracle11g数据库中建立了用于数据挖掘的数据仓库，对数据进行了预处理。在此基础上，用SPSS Clementine11.1数据挖掘工具对企业2008年的客户理赔数据进行了数据挖掘，对结果进行了分析，为企业改善客户关系提出了一些建议，具有一些实用性。 本文从理论上和实验上证明了本文所改进算法的有效性，并将决策树分类技术应用到了物流企业的客户分析中，拓宽了数据挖掘的应用领域。 关键词：物流企业，客户分析，ID3算法，粗糙集，属性依赖度 ABSTRACT In recent years，accompanied with the development of ogistics Industry, more and more third-party logistics enterprises arise and the competition among them become increasingly fierce. Enterprise who wants to have a dominant position in this competition must hold its customers, the premise of which is to fully apprehend the customer’s needs and behaviors. We should apply the data mining technology to analyze the customer’s information which stores in the company’s database. We may find valuable information from these data so as to maintain existing customers and attract new customers, to improve customer’s satisfaction, thus increasing their loyalty and enhance their value and profit. This will become a competitive weapon and play a very important role in strengthening the third logistics enterprises’ competitiveness. In this paper, we mainly do two works. First, the data mining decision tree classification algorithms are studied, and for the defects of ID3 algorithm we propose RID3-1 algorithm. Second, a decision tree classification algorithm is used in the customer claim data of a logistics enterprise; we establish a data warehouse and mine the data, through an analysis of the results we give some suggestions to help the enterprise to improve the