银行信贷管理信息系统的分析与改进研究

【作者】 杨朝晖；

【导师】 王泽兵；

【作者基本信息】 浙江大学 ， 计算机应用技术， 2005， 硕士

【摘要】 近几年，信贷风险已渗透到银行的每一信贷经营之中，严重影响到银行的生存和发展。出于控制信贷风险等业务上的紧迫需求，建设银行开发出了信贷管理信息系统用于建行的信贷管理。 本文在对数据仓库技术、OLAP进行研究的基础上，分析了建行信贷管理信息系统的系统结构，着重阐述了数据仓库应用方案。同时，也研究了数据挖掘技术在信贷管理信息系统中的应用，以达到改进系统的信贷管理效能的目的。具体地，完成了如下实例研究：①利用判定树算法对信用度好的个人信贷客户进行归纳分类模型的建立。②将两种不同的聚类算法用于信贷风险评价指标权重分析，使得信贷风险评价能做到更加合理有效。另外，在自己多年的实践过程中，发现建行现用的这套系统在数据分布模型上存在问题，因而提出了一种两层数据分布的改进模型，让数据适度集中，从而有效解决现有系统所存在的安全方面等的问题。并在最后一章对前面的工作做了一个总结，提出进一步展望。更多还原

【Abstract】 During the past several years, credit risks have penetrated into every corner of the credit businesses, thus seriously impacts the existence and development of banks. For the urgent purpose of the credit businesses, such as credit risk management, China Construction Bank has developed the Credit Management Information System.This document discusses not only the structure of the system and those technologies that it uses in detail, but also how the data statistical analysis is realized. And, it also studies the possibility of using data mining technologies in the system, just to improve the system’ s capability of management. Then it completes the case researches below with data mining technologies:(1)With decision-tree algorithm , it creates a model which classifies the credit degrees of personal loan customers. (2) With two different cluster algorithms , it has analyzed the index weight of credit risk assessment, which can lead credit risk assessment to become more reasonable and effective. Furthermore, with my several experience of managing the system, I have discovered the data location model does exist some very obvious shortcomings. So, this document proposes a new two-level data location model that requires the data to be converged relatively to solve the problems such as security that exists in the system effectively. The last Chapter summarizes the chapters discussed before.更多还原