基于大数据的银行客户贡献度与交易行为模式研究

发布时间：2018-06-08 05:39

本文选题：客户群分析 + 决策树　；参考：《辽宁科技大学》2016年硕士论文

【摘要】：银行业是我国金融体系的主体,当前金融业市场化改革加速,多主体、多层次、多领域的综合性金融市场格局日渐形成。民间金融、产业资本、互联网金融新业态的跨界竞争正在加剧形成新的金融生态体系。大数据、云计算、互联网、物联网技术的飞速发展,极大改变了金融的渠道和工具。在新技术的驱动下,客户的商业模式、发展模式、金融需求都将发生持续而深刻的变革。银行服务涵盖范围广,本文根据某银行的二家分行一年的交易数据,在大数据理论与数据挖掘方法,考虑决策树C5.0算法处理离散化数据的优势,采用决策树C5.0算法对顾客贡献度和交易行为进行建模,并应用SPSS公司的Modeler软件对海量数据进行了处理。研究将准备好的数据利用C5.0分类方法按照客户贡献度区分为几个子群,观察贡献度与其他属性的相关性。在探讨了客户个人数据与总贡献度的关系,以C5.0与CRT分类方法探讨高贡献度客群的交易行为。研究显示,在高贡献度的研究中,(1)高、中贡献度的顾客群男性占较多数。反之,低贡献度顾客群女性较多;(2)30-40为主要贡献顾客群;(3)60岁以上为主要低贡献度顾客群,其趋势与性别无关。而高贡献客户的行为特征研究中,高贡献度顾客群有下列特征:(1)不常使用定期存款者;(2)中度使用贷款业务,代收业务者;(3)高度使用票据、电话银行、网络银行与汇款者。本研究提出一个数据挖掘研究银行客户交易数据的行为的发展模式。可以获得统计学等传统方法无法获得的知识,有助于提高银行的决策能力。
[Abstract]:The banking industry is the main body of the financial system in our country. At present, the market-oriented reform of the financial industry is accelerated, and the comprehensive financial market pattern of multi-subject, multi-level and multi-domain is gradually formed. The cross-border competition among private finance, industrial capital and internet finance is intensifying to form a new financial ecological system. The rapid development of big data, cloud computing, Internet of things and Internet of things technology has greatly changed the financial channels and tools. Driven by new technology, the customer's business model, development model, and financial needs will undergo sustained and profound changes. Banking services cover a wide range. According to the transaction data of two branches of a certain bank in one year, this paper considers the advantage of decision tree C5.0 algorithm in dealing with discrete data in big data theory and data mining method. The decision tree C5.0 algorithm is used to model the customer contribution and transaction behavior, and the Modeler software of SPSS is used to process the massive data. The prepared data are classified into several subgroups according to the customer contribution by using C5.0 classification method, and the correlation between the contribution and other attributes is observed. In this paper, the relationship between customer personal data and total contribution is discussed, and the transaction behavior of high contribution group is discussed by using C5.0 and CRT classification method. The study showed that in the study of high contribution degree, the customer group with medium contribution rate was higher than that of male group. On the other hand, women with low contribution degree were more likely to be the main contribution customers group than 30 to 40 years old, and the trend was not related to gender. In the study of behavior characteristics of high-contribution customers, the following characteristics of high-contribution customer groups are as follows: (1) those who do not often use time deposits (2) moderate use of loans, and (3) customers with high use of bills, telephone banks, network banks and remittance remitting agents. In this study, a data mining model is proposed to study the behavior of bank customer transaction data. Knowledge that can not be obtained by traditional methods such as statistics can help banks to improve their ability to make decisions.
【学位授予单位】：辽宁科技大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：F832;TP311.13

【相似文献】