论文标题:数据挖掘在信用卡客户细分与目标营销方面的应用研究
Research on the Applications of Data Mining in Credit Card Customers Subdivision and Target Marketing
论文作者
论文导师 何建敏,论文学位 硕士,论文专业 管理科学与工程
论文单位 东南大学,点击次数 206,论文页数 59页File Size842K
2006-04-03论文网 http://www.lw23.com/lunwen_975992/
credit card; data mining; customers subdivision; association rule; target marketing
随着人们消费习惯的改变,信用卡业务的巨大发展潜力正在逐步发挥出来,并将成为银行利润的重要组成部分,但竞争也日趋激烈。保留优质客户、提升潜在优质客户,是提高银行竞争力的关键。银行需要更好地了解客户的信息,并将这种信息转变为“知识”,从而更好地为客户提供高质量的个性化服务。数据挖掘技术能够从海量的信用卡业务存储数据中发现一些未知的、有价值的规律,无疑将成为银行提供个性化的信用卡服务的强有力工具。 本文以银行信用卡客户交易数据和客户的人口统计数据等作为研究对象,利用数据挖掘的理论、技术和方法挖掘银行信用卡信息中的知识,并利用这些知识服务于信用卡客户细分与目标营销。本文的主要工作有: 1.针对信用卡的特征,提出了从消费行为和还款行为两方面的客户价值度量的指标RFMRF。 2.对数据挖掘的算法进行了介绍,重点介绍了SOM、FP-tree等算法,并用改进的SOM算法对信用卡客户进行细分,将客户按照贡献度的大小分为优质客户、明星客户、大众客户以及利润消耗型客户。 3.利用FP-tree算法挖掘出各种类型客户的特征,并且从动态的角度来挖掘出易于变化的客户(从一种客户类型转变为另一种客户类型)特征属性,有助于新客户的搜寻和为现有的客户提供差异化的营销服务。 4.根据研究的结果提出银行信用卡业务的营销策略,为未来信用卡的有效营销提供了参考。
Along with the changing of people’s consuming habits, there is a huge growing potential in credit card business in China, but the competition becomes more and more fierce. To hold high-quality customers and upgrade potential high-quality customers are keys for banks to improve competitiveness. Banks can provide higher quality individuation service for customers by getting well known of customers’information and changing this information to knowledge. The technology of data mining refers to finding out some unknown and valuable rules from a large number of credit card transaction data and it is a strong tool for banks to provide individuated service of credit cards. With bank’s credit card transaction data and customers’vital statistics data as research object, this paper uses the theory, skills and methods of data mining to acquire knowledge hiding under the information of credit cards and serve for credit card marketing. Its main tasks are following: 1. Aimed at characteristics of credit card, it puts forward the index RFMRF for measuring customers’value from the behaviors of both consumption and paying back. 2. It introduces the arithmetic of data mining and puts emphases on SOM and FP-tree. Then it subdivides customers into three kinds—high-quality, star and common customers according to their contributions by improved SOM arithmetic. 3. It obtains the characters of three kinds of customers by FP-tree and dynamically mines the attributes of customers who are easy to change, so as to search new customers and provide different marketing service for old customers. 4. Based on the research result, it brings forward sales and marketing strategy for banks’credit card business and gives a reference for effective sale and marketing of credit cards in future.
|
