Research On Green Credit Credit Evaluation Based On Convolutional Neural Network

As a new bank credit product,green credit must strengthen the research on credit evaluation of green credit in order to ensure its wider and smooth development and play its true role.For a long time,the credit evaluation system of commercial banks in China is not perfect,and there is no efficient green credit evaluation model to control and measure the default risk of enterprises.Nowadays,with the continuous development of information technology,artificial intelligence methods and machine learning models are applied to solve the problem of credit evaluation.This paper attempts to do some research and research on green credit evaluation with the convolutional neural network model that has not been applied in the field of credit evaluation.This paper focuses on the evaluation of green credit,based on sustainable financial theory,corporate social responsibility theory,equatorial principle theory and bank credit evaluation theory,and summarizes the relevant research on green credit and its evaluation management at home and abroad.So,the connotation and development status of green credit evaluation indicators are summarized and analyzed.On this basis,the selection of indicators and their causes are analyzed and elaborated.Then,based on the relevant theoretical analysis of the previous papers,a green credit evaluation index system including traditional financial indicators and ESG(Environment,Society,and Governance)development indicators is constructed.After that,the data from 247 listed companies in Shanghai Stock Exchange and Shenzhen Stock Exchange were selected,and the appropriate commercial bank green credit evaluation model was established by using support vector machine,BP neural network and convolutional neural network,and the results were compared and analyzed..The research results show that compared with the traditional credit evaluation model that only contains financial data,the green credit evaluation system that introduces ESG development indicators can better reflect the credit level of enterprises.At the same time,the paper also finds that the convolutional neural network model has better classification and prediction ability and better accuracy than the support vector machine and BP neural network classification without human intervention.