基于AEN-logistic模型的中小企业信用风险评估  

Credit Risk Assessment for SMEs Based on A EN-logistic Model

作　　者：李淑锦[1] 吴王滢 LI Shujin;WU Wangying(School of Economics,Hangzhou Dianzi Univrsity,Hangzhou 310018,Zhejiang,China)

机构地区：[1]杭州电子科技大学经济学院,浙江杭州310018

出　　处：《信息与管理研究》2021年第4期27-36,共10页Journal of Information and Management

基　　金：国家社会科学基金项目(17BJY233)。

摘　　要：从偿债能力、盈利能力及非财务指标如股利分配等角度选择具有代表性的指标构建中小企业的信用风险评估指标体系,将弹性网络[Elastic Net(EN)]和对齐加权[Adaptive Elastic Net(AEN)]与logistic结合,创新性地建立EN-logistic和AEN-logistic模型来预测中小企业的违约概率。实证结果证明EN-logistic和AEN-logistic模型均大大提高了模型的预测精度,特别是AEN-logistic模型,预测准确度高达90%,说明AEN-logistic模型在中小企业信用风险评估中是有效的。Representative indexes are selected from such perspectives as solvency,profitability,and non-financial indicators(e.g.dividend distribution)to construct a credit risk assessment index system for small and medium-sized enterprises(SMEs).The elastic network Elastic Net and alignment weighted Adaptive Elastic Net are combined with logistic to establish the EN-logistic and AEN-logistic models so as to predict the default probability of SMEs.The empirical results show that both the EN-logistic and AEN-logistic models greatly improve the prediction accuracy of the model,and especially the prediction accuracy of the AEN-logistic model reaches 90,which indicates that the AEN logistic model is effective in the credit risk assessment of SMEs.

关 键 词：信用风险 弹性网络 对齐加权 LOGISTIC模型 中小企业 

分 类 号：F830[经济管理—金融学]