摘要
本文以广东省各个行政部门的企业数据为基础,研究了证据权重在企业信用风险等级评估中的应用,并且根据国家“企业信用风险分类评价标准”系统建立了企业信用评价指标,成功地将证据权重逻辑回归算法应用到真实的企业信用数据集,建立企业信用风险评估模型,使得监管部门能够准确地掌握企业的信用情况。通过与经典的逻辑算法比较,验证了该方法的有效性。
Based on the enterprise data of various administrative departments in Guangdong Province,this paper studies the application of weight of evidence in the assessment of enterprise credit risk levels,and establishes enterprise credit evaluation indicators based on the national"Enterprise Credit Risk Classif ication and Evaluation Standards".The weight-of-evidence logistic regression algorithm is applied to real enterprise credit data sets to establish enterprise credit risk assessment model,enabling regulators to accurately grasp the enterprise credit situation.Compared with the classical logic algorithm,the validity of the method is verif ied.
作者
危明铸
沈凤山
袁峰
麦伟杰
WEI Mingzhu;SHEN Fengshan;YUAN Feng;MAI Weijie(Institute of Software Application Technology,Guangzhou&Chinese Academy of Sciences,Guangzhou,Guangdong Province,511458 China)
出处
《科技创新导报》
2020年第29期137-143,共7页
Science and Technology Innovation Herald
基金
广东省“芯片、软件与计算”(软件与计算类)重大专项(项目编号:2020B010166005)。
关键词
证据权重
逻辑回归
信用风险
企业信用
Weight of evidence
Logistic regression
Credit risk
Credit of enterprise