基于聚类技术的银行客户行为变化研究

发布时间：2018-04-18 11:03

本文选题：数据挖掘 + 分类　；参考：《中南大学》2013年硕士论文

【摘要】：近些年来,客户关系管理受到学术界和从业者们越来越多的关注,这是因为市场竞争环境和技术的进步,使得客户行为的变得很难预测。客户关系管理是一种新的营销方法,使得银行将自己的业务由从以产品为中心转移到以客户为中心。依据客户各自的特征客户计划进行确定的、有意义的分类的,实现这一转换。客户细分是客户关系管理(CRM)的主要核心功能之一,而细分的依据主要有以下几个方面：价值、社会人口统计、忠诚度、需要/态度和行为。本文着重于客户行为划分。尽管很多的研究证明了CRM策略的重要性,但是对客户兴趣和重要性的增长的理解还是不够的。因此,本论文的目的是提供一个综合性的框架来帮助研究人员集中于客户关系管理的研究,同时,来帮助相关行业人员和市场营销人员通过数据挖掘方法来对客户进行分类,从而成功实现客户关系管理的目的。该框架建立在客户关系管理和数据挖掘方法体系的基础上,并且提供了一个系统方法,该方法实现了客户分类中CRM集成到企业市场营销策略中的实现。为了评估银行客户分类的CRM框架,通过对银行信用卡的消费记录来对客户的行为进行分类,这种分类基于他们的采购行为尤其是他们即将采购的产品进行划分。通过一个两步的聚类算法把这些客户划分成五部分。我们通过一个可以提供更好的模型和74.8%精确度的CHAID决策树来描述客户的分类。
[Abstract]:In recent years, customer relationship management (CRM) has attracted more and more attention from academia and practitioners. This is due to the development of market competition and technology, which makes customer behavior difficult to predict.Customer relationship management (CRM) is a new marketing method that enables banks to shift their business from product-centric to customer-centric.A defined, meaningful categorization of customer plans according to their own characteristics, to achieve this conversion.Customer segmentation is one of the core functions of CRM (customer relationship Management), and the subdivision is based on the following aspects: value, social demography, loyalty, need / attitude and behavior.This paper focuses on the division of customer behavior.Although many studies have proved the importance of CRM strategy, the growing interest and importance of customers is not enough.Therefore, the purpose of this paper is to provide a comprehensive framework to help researchers focus on customer relationship management, and to help relevant industry personnel and marketers classify customers through data mining methods.In order to successfully achieve the purpose of customer relationship management.The framework is based on the system of CRM and data mining, and provides a systematic approach, which realizes the integration of CRM into enterprise marketing strategy in customer classification.In order to evaluate the CRM framework of bank customer classification, the customer behavior is classified through the consumption record of bank credit card, which is based on their purchasing behavior, especially the products they are about to purchase.These customers are divided into five parts by a two-step clustering algorithm.We describe customer classification through a CHAID decision tree that provides a better model and 74.8% accuracy.
【学位授予单位】：中南大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP311.13

【共引文献】