Application of Normal Multivariate Binary Scale Mixing in Regression Model
Download as PDF
Credit risk is the most important risk faced by commercial banks, and it has a crucial impact on the bank’s continued sound operation. How to effectively manage credit risk on the basis of accurate measurement is a very challenging issue facing commercial banks. This article uses public financial data of listed companies to establish a Cox credit risk normal multivariate binary scale mixed regression model. The selected indicators have high dimensionality and strong correlation, and contain a lot of redundant information. How to select the covariates that really affect the response normal multivariate from the many information for modeling is very necessary. In this paper, by analyzing the application scope, data requirements, prediction accuracy and stability of each model, this paper finds that the Logistic regression model is more suitable for the actual situation in China in terms of input data and premise assumptions. Therefore, this paper finally chose the Logisti regression method to establish credit Risk measurement model, and an empirical test of the normal multivariate binary scale for the effect of the regression model. When using the logistic regression method to establish the credit risk measurement, this paper proposes a method of objectively selecting the input parameters of the logistic model. The experimental results show that the normal multivariate binary scale is used to distinguish the normal companies above 80%. However, there is still a significant gap compared with the logistic model's judgment accuracy of training samples above 90%.
Binary Scale of Positive Polymorphic Variables; Logistic Regression Model; Credit Risk; Factor Analysis