论文部分内容阅读
关于信用风险评价问题至今已经做了很多研究,各种信用评价模型与方法也已被开发。但是这些模型与方法几乎都是基于财务数据、股票价格或风险调研机构发表的各种调查结果。因为几乎所有的中小企业的财务数据都是非公开的,至今开发的信用评价模型与方法都不免成为无米之炊。为此,本文提出了一个新的途径,只需要根据销售额、顾客付款额、拖欠款额等日常业务处理数据来评价顾客企业的信用度。本文提出一个应用Bagging方法评价顾客信用的系统,其目的在于解决由于异常顾客数比正常顾客要少很多而带来的问题,提高分辨异常顾客的能力。本文所提出的信用评价系统将应用到一个实际企业的信用评价问题中,借此来验证系统的性能和效果。
Much research has been done on credit risk assessment so far, and various credit evaluation models and methods have also been developed. However, almost all of these models and methods are based on financial data, stock prices or various surveys published by risk research organizations. Because almost all the financial data of SMEs are non-public, the credit evaluation models and methods developed so far are inevitably short of money. To this end, this paper presents a new way, which only needs to evaluate the customer’s credit according to daily business processing data such as sales, customer payment, arrears and so on. This paper presents a system using Bagging method to evaluate customer credit. The purpose of this paper is to solve the problem brought by the much smaller number of abnormal customers than normal customers and to improve the ability to distinguish abnormal customers. The credit evaluation system proposed in this paper will be applied to the credit evaluation of a real enterprise to verify the performance and effectiveness of the system.