Credit assessment for listed companies based on GA-BP model

This paper first constructed a financial index system for credit assessment of listed companies, then analyzed the defects of BP neural network model and found out it is easy to fall into the defect of local optical solution. By introducing the revised connection weight value of genetic algorithms, constructed GA-BP neural network model, and selected 80 companies listed from 2004 to 2007 as training samples, another 80 listed companies as test samples, to evaluate after screening indicators. The result shows high accuracy on assessment by using this model.