Research On The Artificial Intelligence Credit Decision Model Of Commercial Banks

As we all know,commercial banks in China are the engines of the national economy and have a very critical and important role.Asset quality safety is the topmost priority for a commercial bank.Credit risk management is the core part of credit risk management.Therefore,a scientific and effective credit risk decision-making mechanism is of great significance to commercial banks.With the continuous development of social economy,the credit decision-making mechanism of the commercial banks needs not only to control the risks effectively,but also requires the substantial increase of efficiency.Therefore,the auxiliary credit decision-making model has also emerged.The model designed in this dissertation is based on financial theory,financial risk management theory and artificial intelligence theory.Qualitative and quantitative analysis method as well as integrated methods are adopted to study the credit decision-making of commercial banks.Many researchers,both nationally and abroad,have used a variety of engineering theories or techniques to construct a wide range of risk management models,but most of these studies are based on static modeling of past time samples of data.The model built in this thesis aims to solve the defect that dynamic modeling lacks a dynamic updating mechanism.This model firstly deals with the report data with the powerful solvency and ability to the practical problems of artificial neural network system,and quantitatively analyzes the possibility of loan repayment;Secondly,the expert system is used to qualitatively analyze the possibility of loan repayment in terms of non-financial and credit support;Then,according to the five principles of comprehensiveness,generality,sensitivity,measurability and priori,the author establishes the index system of bank credit decision modeling;Finally,an empirical research is made on the dynamic modeling of a commercial banks'credit decision based on incremental SVM of branch-and-bilk batch to realize a dynamic update mechanism of bank credit decision.The results of the empirical study show that the model constructed in this paper has decision-making ability under real conditions and can dynamically adjust the model according to the increase of new samples over time,thus it can keep a good ability to to update the model itself in order to provide a decision support to the bank credit decision-making.In general,the practical research in this paper can provide some practical ideas and examples for the theoretical research of credit decision-making system in commercial banks.