Font Size: a A A

ANN Model Appling To Personal Credit Risk Management

Posted on:2009-10-08Degree:MasterType:Thesis
Country:ChinaCandidate:J CaoFull Text:PDF
GTID:2189360245974623Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
With the reform and opening to the world, Domestic economic and society developed rapidly. Also Personal Credit got a great progress in 21 century: Because the price of house was Soaring, more and more people cannot buy it in cash, so what they can do is get Mortgage loans from bank and other Financial Institution; Education loans make lots of students have chance to learn in campus; Personal consumption loans had been an important method to Promoting Domestic demands and growing of economic. They bring bank Personal Credit Risk, how to manage it has been an difficult problem to be solved.Traditional Credit Risk assessment methods used to analysis Enterprise Credit Risk(ECR). There are many defects when these methods applied to Personal Credit Risk (PCR). Such as subjective factors when some methods applied to PCR, and sometimes the results have volatility, all these reasons Impact the accuracy of the results. Through learning the methods of ECR Assessment and analysis the Special feature of PCR, it find that ANN Model can bring down the Probability of the defects and give an accuracy assessment result relatively.In this paper, there is an introduction of theory related to Credit Risk, introducing the management theory,research methods and development of PCR in our country. Then, it analysis the factor which impact PCR and establish the indicators system under the domestic economic and society conditions. Last step, there preprograms the scores of the indicators using Factor Analysis, then choose BP algorithm to establish the ANN Model which is used to computing the results of PCR. Bank can make reasonable decisions and manage the PCR effectively according to it.
Keywords/Search Tags:Commercial Credit, Credit Risk, Factor analysis, BP Artificial Neural Networks
PDF Full Text Request
Related items