聚类分析在银行个人贷款客户细分中的应用.pdfVIP

摘 要 2006 年底，我国经济金融结束了加入WTO 的过渡期，全面融入国际体 系。外资银行也乘势大举进军国内市场，各类国内外金融机构也如雨后春笋 般涌现。银行业的竞争空前激烈。而2008 年的金融危机又给银行业带来了 巨大灾难。为了避免控制银行业的风险，各国政府都不断的加强对银行的监 管。面对激烈的市场竞争和严厉的金融监管，商业银行要想生存就必须在精 细化管理上下功夫提升自身的盈利水平。而进行精细化管理，客户细分是基 础工作。 本文运用聚类分析对商业银行客户信息进行聚类分析，并将聚类分析结 果按照客户利润贡献矩阵理论进行客户分类，最终将客户分为优质客户、重 要客户、一般客户、潜力客户和亏损客户。并对每类客户制定相应的营销管 理方案。本文设计了具体的聚类实验步骤，并通过使用SPSS 统计分析工具 的K-means 分类功能来对银行客户信息进行细分。实验首先对数进行了预处 理。然后设计预聚类，将噪声数据去除，并得到初始聚类中心后备集。接着 通过对K 进行多次不同取值，并绘出误差曲线图，用以确定最终的K 取值和 K 个初始聚类中心。在K-means 聚类参数都确定后，得到最终的聚类结果。 按照聚类结果分析的客户现实和潜在利润贡献值，并按照利润贡献矩阵，将 聚类结果归为客户细分结果。研究结果在对指导银行个性化产品开发及客户 分类管理有一定的实际意义。 : FTP K-Means 关键词 数据挖掘 客户细分 I - - Abstract The transition of Chinese economy after our country joined WTO drew to a close at the end of 2006 and Chinese economy has already fit into the international economic system. Large numbers of foreign banks have got into the domestic market, as well as various national and international financial institutions. The competition in the banking industry has continually heat up. The financial crisis in 2008 brought the banking industry great disaster. To avoid such risks, governments have improved supervision of banks. Commercial banks must refine the management to enhance their level of profitability in order to survive in the industry competition. And the refined management and customer segmentation is the basis for work. In this thesis, customer information of commercial banks is analyzed with the cluster analysis. Then the results can be classified into 5 types including high-quality customers, important customers, ordinary customers, potential customers and deficit customers according to the customer profit distribution matrix. The marketing plans should