| This paper studies the credit risk measurement of software and information industry.Credit risk involves the most extensive range of financial risks.It needs to be prevented and managed urgently.Software and information industry is a new strategic industry in China,which is also a medium and long-term development project.But there are many problems such as small scale and difficult financing.Therefore,it is so important to establish a credit risk measurement model for software and information industry.Firstly,the KMV model is improved by looking for the default points of listed companies in software and information technology industry.In this paper,the default point coefficient combination of(1,0.5)in the traditional KMV model is abandoned.This paper adopts dynamic particle swarm optimization algorithm,makes a linear change strategy for weight and learning factor,and transforms this problem into a maximum problem of the default distance gap between labeled *ST and unlabeled companies.The optimal coefficient combination of default point is found as(1.2171,8.2559)by searching.Then,the improved KMV model is used to measure the credit risk of the listed companies.By estimating and setting the five parameters of the model,the asset value and its volatility of the listed company from software and information industry can be obtained.Next,with the default point,the expected default distance of listed companies can be calculated.And,a new credit risk forecasting model is constructed to forecast the principal credit rating of the company.Finally,the improved BP-KMV model is established to solve the problem of credit risk measurement of unlisted companies effectively.BP neural network is introduced into the KMV model.The input item is financial data from company,and the output item is its asset value and its volatility.The improved KMV model is used to solve the expected default distance of unlisted companies.Compared with the rating of credit risk issued by related rating agencies,it is found that this improved model is rational and applicable,and can overcome the lack of equity data for application to credit risk measurement of unlisted companies. |
