Research On Credit Risk Evaluation Of Credit Card In Commercial Banks

With the rising living standards, residents’consumption habits have also be changed, and the concept of advance consumption is growing in popularity,which has promoted the rapid development of credit card in commercial banks. Credit card is a kind of business without any collateral, however, there are many uncertain factors in economic life, which lead to the increasingly default rates on the credit card of commercial banks. Especially in recent years, default rates on credit card is increasing even faster. So commercial banks should put a high premium on prevention. In addition, after many years of accumulation, there is a large number of historical data in credit card system.However they are drowning in mass data but starved of information. Secondly, many foreign advanced banks have explored many effective methods of credit risk assessment system, and these methods have applied to the issuing agencies. Because of the lack of experience in risk assessment of credit card in China,we have a lot of wok to do,compared with the high income of the credit card in foreign banks. Experience shows that the high profit of credit card is built on the basis of better control of the credit risk of credit card. Therefore, only by making full use of credit card business data, being combined with the credit card credit risk management theory, using the appropriate credit card credit risk assessment methods, establishing a good card credit risk assessment model, can commercial banks achieve high efficiency of credit risk management of credit.This paper firstly gives a brief introduction of organization,development modelthe,and current situation of credit card, presents three main risks of it, and points out that the credit risk is the major risk of the credit card, which is should be controlled urgently.And then analyses reasons of the risk from the aspects of cardholder, card issuers and others. After analyzing and comparing different credit risk assessment models, such as Logistic Regression and Decision Tree,finally, Random Forest is chosen as the empirical analysis model in flowing empirical analysis.The data for empirical analysisis is from the Internet and the main variables include eage, marriage, education, occupation, funds in the account balance and so on. The results of empirical show that balance,age,dayjob,loan,housing have a significant impact on credit card evaluation. In addition,Decision Tree can provide intuitive decision support to decision makers, Logistic Regression can predict the default probability.However,in model fitting goodness and generalization error, Random Forestdoes much better than them which provides solid theoretical support for the assessment of credit risk of credit cards.Finally, according to empirical analysis and other related theories, this paper provides the theoretical basis and technical support for commercial banks to choose appropriate credit card credit risk evaluation methods, strengthen the bank’s internal information construction, establish and improve the external credit rating system to strengthen the credit risk prevention of credit card.