银行个人客户挖掘与营销系统

发布时间：2018-01-02 13:03

本文关键词：银行个人客户挖掘与营销系统　出处：《西安电子科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：个人客户是银行的主要客户群体。随着银行业竞争的日益激烈,互联网金融对银行的冲击越来越大,针对不同客户特点,为客户提供不同的服务和推荐客户更感兴趣的产品已成为银行制胜的法宝之一。银行通过多年联机事务系统的运营积累了大量的数据,缺少一套通过对数据进行分析和挖掘,帮助业务营销的系统。为解决此问题,分行启动了个人客户挖掘与营销系统的建设,该系统通过对个人客户相关信息的整理、分析和挖掘,为业务营销提供支持,同时为管理机构提供数据统计等功能和为业务决策提供帮助。本文首先讨论了数据仓库、数据集市和相关技术,分析了传统数据库和数据仓库的差异,探讨了数据仓库的分类和系统结构。然后,从用户的需求入手使用面向对象的分析技术通过类图、用例模型、顺序图等方法,分析了系统实现需要完成的主要工作。分析时还从非功能性需求方面对系统的性能、数据安全、数据备份与恢复策略进行了考虑。对总行和分行的数据进行分析,明确了建立数据集市所需的数据源。其后,根据数据仓库建模技术对系统的数据层进行了概念模型、逻辑模型和物理模型的三个层级的抽象和分析,最终设计了省分行从属型数据集市的“两级、三层”的系统结构,其中“两层”是指数据的存储的划分,“三层”是指系统的实现结构划分。随后,根据系统分析的结果从数据抽取/转换/装载、数据挖掘、数据展现等方面对系统实现的关键模块进行了设计。系统通过基于关联规则挖掘实现产品关系的规则库,当对某一具体客户进行营销时,系统根据客户已持有的产品在规则库中进行匹配,将匹配的结果推荐给客户。通过数据挖掘的手段提高产品推介时的成功率,进一步提升用户感受。最后,对系统实现过程中所使用的软硬件环境进行了简单的说明,展示了系统运行后的各个模块的实际运行效果和在数据安全方面所做的工作。通过本次系统设计到实现,初步探索了一套在国内银行业有效的分行级数据集市建模方法。业务部门在对系统验收后反馈:系统已基本满足业务需求,完成了既定的目标。通过对系统的分析后续还可以进一步的对数据进行挖掘,同时增加更多维度的数据分析,让数据在业务营销中发挥更大的作用。本次系统的实现也为后续面向其他主题分析系统的建设积累了经验,为逐步完成企业级数据仓库奠定了良好的基础。
[Abstract]:Individual customers is the main customer groups of banks. The banking industry increasingly fierce competition, the Internet financial impact on banks is more and more big, for different customers, provide different services and recommend that customers are more interested in the products to customers has become one of the magic weapon for winning bank. The bank has accumulated a large amount of data through years of online transaction the lack of a system operation, through the analysis and the data mining system, to help business marketing. In order to solve this problem, the branch started the construction of individual customer mining and marketing system, the system through the relevant information of the individual customer collation, analysis and mining, provide support for business marketing, data statistics and other functions for the management of institutions and provide help for business decisions at the same time. This paper discusses data warehouse, data mart and related technology, the traditional database and data analysis Warehouse differences, discusses the classification and system structure of data warehouse. Then, from the user's demand of using the technology of object-oriented analysis by class diagram, use case model, sequence diagram analysis method, the system needs to complete the work. The analysis of system performance, from the non functional requirements of data security, data backup and recovery strategies were considered. The head office and the branch of the data analysis, clearly needed to establish the source of data for the data marts. Subsequently, according to the data of data warehouse modeling technology on the system layer of the conceptual model, three levels of abstraction and analysis of logical model and physical model, the final design the branch of the subordinate data mart "two levels, three layer system structure", the "two layer" refers to the division of data storage, "three" refers to the realization of the system structure is classified. Then, root According to the results from the analysis of the data extraction system of ETL, data mining, data display and other aspects of the design of the key modules in this system. The system through the rule base to achieve product relationship based on association rule mining, when marketing to a specific customer, the customer has to hold the system according to the matching rules in products in the library, the matching results to recommend to customers. By means of data mining to improve the success rate of product promotion, to further enhance the user experience. Finally, the system hardware and software environment in the process of using a simple description, showing the effect of actual operation of each module of the system after running and done in data security. This system is designed to achieve, explore a set of effective in the domestic banking branch level data mart modeling method. Business Department of the Department of The acceptance of feedback: the system can satisfy the needs of the business, to complete the established goals. Through the analysis of the system can also carry out further follow-up of data mining, while increasing the analysis more dimensions of data, make data play a more important role in business marketing. Implementation of this system for the follow-up for other topics analysis of accumulated experience of system construction, and laid a good foundation for the gradual completion of enterprise data warehouse.

【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP311.13

【参考文献】