摘要
为了探寻以政治认同为代表的非经济因素对信用风险产生的影响,将非经济因素的相关变量纳入金融分析框架的思路。整合经济因素及非经济因素数据,通过 BP神经网络和循环神经网络LSTM对问题进行建模分析的研究方法,研究表明,企业高管的政治认同程度对企业的信用风险有显著影响;企业高管的政治认同程度越高,信用风险的发生概率相对越低;加入企业高管政治认同因素的相关数据进行分析能够有效提高对信用风险的预测能力。
To explore the potentially great influence of non-economic factors represented by political identity on credit risk,the correlated variables of non-economic factors were incorporated into the financial analysis framework.This paper selects the BP and loop LSTM neural network which are equipped with biology evolution characteristics and strong fitting ability to do the modeling analysis.The research results are as follows,the degree of political identification of corporate executives has a significant impact on corporate credit risk;the higher the political identification degree of corporate executives is,the lower the probability of credit risk appears;the analysis of data integrated with political identity factors of corporate executives can effectively improve the ability to predict the credit risk.
出处
《科学决策》
CSSCI
2022年第10期81-94,共14页
Scientific Decision Making
关键词
政治认同
违约风险
神经网络模型
political identity
default risk
neural network model
