有序多分类离散选择模型的信用评级与风险预警研究——以国内20家上市企业为例

信用风险评估模型的选择直接影响着对上市企业信用等级的评估。本文在建立上市企业信用等级评估指标体系的基础上,采用语言信息处理技术,通过因子分析法提取上市企业信用等级评估的潜在变量,建立有序多分类离散选择模型,给出了信用等级判别分析计算步骤,并以中国境内20家上市企业为例进行了实证分析。以上设计运用定量和定性指标分析相结合的方式,有效地将语言评价信息进行集结和利用,解决了上市企业中风险级别可以排序、定性的问题。可促进上市企业信用评级制度的完善,实现对上市企业信用规范管理的目标,从而建立健全全社会信用评价体系。

Ordered Discrete Choice Model-Based Credit Rating and Risk Warning: Illustrated by an Example of Twenty Domestic Listed Companies

In order to solve the listed company credit evaluation and early-warning management,an analysis approach based on multi-classification discrete choice model is proposed.On the basis of building a listed company credit evaluation index system,applying the linguistic information processing method,the panel valuation data is aggregated to obtain the integrated evaluation values.The potential variables are extracted by factor analysis method,an ordered logistic model is put forward,and the calculation analysis steps are given.This research is illustrated by an example of twenty domestic listed companies.The research indicates that the linguistic information can be effectively aggregated with less lost by this method,the issues that the classes of listed company credit are ordinal,and approach will promote the improvement of the credit rating system of listed companies,and achieve the target on the standardized management of listed company credit.