Research On Customer Credit Card Forecast Based On AHP-logistic Model

As we all know,in the economic field,risks and benefits are a pair of coexisting contradictions.Whether in financial products or in banking,they all show high risks,high returns,low risks and low returns.Since the birth of the first credit card in 1952,the credit card has included the double-layered features of financial products and banking.Credit cards can allow us to spend the next month's money in our personal and corporate terms.Dream;in the card issuer,that is,the bank,you can realize the current money and receive the benefits of tomorrow.The credit card business is different from the usual financial products and banking business.First,it has the financial attributes of financial products and banking;secondly,it is an important payment method from the paper currency to the electronic money evolution process;finally,it Social attributes that are used by a wide range of social groups and different social fields.The proportion of today's credit card business in China's banking business,the number of card issuance and the popular population are all expanded with the day.This is inseparable from the high income that it can bring to the card issuer,but the risk is also Come along with it.The risk management of the credit card business is a problem that the card issuer must face and properly solve.The content of this paper is exactly in line with it.How to judge and predict the customer's monthly repayment default according to the customer's monthly and previous credit card repayment and consumption situation.Probability,this article is to convert this probability into a concrete score,which is used by the issuer to stop the loss in time or to increase the customer credit limit.The research topic in this paper is based on the known customer gender,age,education level,marital status,credit card consumption in the past 6 months,previous payment and credit card repayment,etc.,and the prediction of the user's monthly default rating.Using logistic regression modeling to study the credit default problem has the advantages of good fitting effect and high stability.Therefore,this paper can also be used for reference.However,this paper does not directly use traditional logistic regression,but introduces customer stratification and analytic hierarchy process.According to the weight of the indicators that are selected into the model,the credit value of each credit card customer is scored,and the customer is divided into high-quality customers,ordinary customers and inferior customers,and the result is introduced into the logistic regression model as a variable.Go and create a logistic regression model based on AHP customer stratification.This kind of improvement makes the prediction of the default of the next month more accurate and effective.The method of this paper has very high referenceability and practicability,and can quickly achieve a more accurate credit score for the customer,so that the card-issuing party can use the credit card for the next month.Effective control.