基于CHAID决策树的信用卡风险管理系统设计
发布时间：2018-11-21 21:16
【摘要】：新锐银行由于信用卡业务经验缺乏,特别是在信用风险控制方面比较薄弱,目前发卡审核流程仅靠人工获取人行征信系统的数据来判断是否批准是远远不够的,还有随着信用卡业务的发展产生了海量的历史交易数据,所以需要对这些数据进行有效地管理和分析。为了避免信用卡信用欺诈和降低信用卡坏账风险,一方面分析客户的基本资料,从源头上控制了信用卡风险。另一方面基于CHAID决策树能够从大量的信用卡交易数据中获取出一些未知的、有价值的规律,而这些规律可以达到预测的目的,无疑降低了商业银行信用卡恶意透支等风险。首先讨论了新锐银行的现状以及其对各种风险管理应对的不足,采用CHAID决策树分析了数据并从中找出有益的信息来降低信用风险是十分有必要的,在明确了该系统设计的基础上,讨论了信用卡的客户管理、卡片申请、消费授权等主要功能,分析了客户信用评分、风险预警等核心流程。然后对客户信用管理、发卡审核、信用欺诈等核心子系统进行详细设计,其中重点讨论了构建信用评分来实现信用等级分类规则以及采用CHAID决策树算法识别信用欺诈等过程,达到信用卡发卡审批和消费交易行为进行预测的应用。最后,对同类的风险管理系统设计进行了比较。
[Abstract]:Due to the lack of experience in credit card business, especially in the area of credit risk control, Xinrui Bank is currently only relying on the manual acquisition of data from the credit system of the people's Bank of China to judge whether approval is approved or not. With the development of credit card business, there is a huge amount of historical transaction data, so it is necessary to manage and analyze these data effectively. In order to avoid credit card credit fraud and reduce credit card bad debt risk, on the one hand, the basic information of customers is analyzed, and credit card risk is controlled from the source. On the other hand, the decision tree based on CHAID can obtain some unknown and valuable rules from a large number of credit card transaction data, and these laws can achieve the purpose of prediction, which undoubtedly reduces the risk of malicious overdraft of credit card in commercial banks. Firstly, the paper discusses the current situation of Xinrui Bank and its shortcomings in dealing with various kinds of risk management. It is necessary to use CHAID decision tree to analyze the data and find out useful information to reduce the credit risk. On the basis of defining the design of the system, this paper discusses the main functions of the credit card, such as customer management, card application, consumption authorization, and analyzes the core processes such as customer credit rating, risk warning and so on. Then the key subsystems such as customer credit management, card issuance, credit fraud and so on are designed in detail, in which the process of constructing credit score to realize credit rating classification rules and using CHAID decision tree algorithm to identify credit fraud are discussed in detail. To achieve credit card approval and consumer transactions to predict the application. Finally, the design of similar risk management system is compared.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.2;TP315
本文编号：2348276
[Abstract]:Due to the lack of experience in credit card business, especially in the area of credit risk control, Xinrui Bank is currently only relying on the manual acquisition of data from the credit system of the people's Bank of China to judge whether approval is approved or not. With the development of credit card business, there is a huge amount of historical transaction data, so it is necessary to manage and analyze these data effectively. In order to avoid credit card credit fraud and reduce credit card bad debt risk, on the one hand, the basic information of customers is analyzed, and credit card risk is controlled from the source. On the other hand, the decision tree based on CHAID can obtain some unknown and valuable rules from a large number of credit card transaction data, and these laws can achieve the purpose of prediction, which undoubtedly reduces the risk of malicious overdraft of credit card in commercial banks. Firstly, the paper discusses the current situation of Xinrui Bank and its shortcomings in dealing with various kinds of risk management. It is necessary to use CHAID decision tree to analyze the data and find out useful information to reduce the credit risk. On the basis of defining the design of the system, this paper discusses the main functions of the credit card, such as customer management, card application, consumption authorization, and analyzes the core processes such as customer credit rating, risk warning and so on. Then the key subsystems such as customer credit management, card issuance, credit fraud and so on are designed in detail, in which the process of constructing credit score to realize credit rating classification rules and using CHAID decision tree algorithm to identify credit fraud are discussed in detail. To achieve credit card approval and consumer transactions to predict the application. Finally, the design of similar risk management system is compared.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.2;TP315
【参考文献】
相关期刊论文 前2条
1 韩彦峰;申琦;;信用卡风险分析及防范[J];商场现代化;2007年03期
2 丁玲娟;;关于信用卡风险防范的研究[J];商场现代化;2009年17期
，本文编号：2348276
本文链接：https://www.wllwen.com/guanlilunwen/bankxd/2348276.html
