基于改进ELECTRE-IN的小微企业信用评估  

Small and Micro Enterprise Credit Evaluation Based on Improved ELECTRE-IN

作　　者：陈伟杰[1] 龚涛 Chen Weijie;Gong Tao(School of Economics and Management, Chongqing Normal University, Chongqing400047, China)

机构地区：[1]重庆师范大学经济管理学院,重庆400047

出　　处：《重庆师范大学学报（社会科学版）》2021年第6期38-50,共13页Journal of Chongqing Normal University(Edition of Social Sciences)

基　　金：2019年度国家自然科学基金青年项目“面向不确定性多源异构数据融合分类方法的小微企业信用评级研究”(71901044);2018年度教育部人文社会科学研究青年基金项目“基于广义模糊软集的多源数据融合决策方法与应用研究”(18YJC630009);2018年度重庆市社会科学规划项目“基于多源数据融合的模糊软集决策方法与应用研究”(2018QNGL28)。

摘　　要：基于小微企业有效评价信息匮乏的难题,以非财务指标为重点,纳入水电费等大数据下体现企业违约风险的因素,构建了小微企业信用评价指标体系。引用ELECTRE-IN消除ELECTRE-Ⅲ中由偏好指数产生的放大效应,完善级别高于关系的构造过程。针对ELECTRE-IN主观性强及排序复杂的问题,首先利用BWM-GiNi组合赋权法确定评价指标的综合权重,并根据样本数据极差确定阈值组合,增强其客观性。其次引用可信度法替代蒸馏算法,简化排序过程。最后利用算例对一批小微企业进行排序,算例结果验证了模型的有效性。Based on the problem of lack of information for effective evaluation of small and micro enterprises,focusing on non-financial indicators, including factors that reflect corporate default risks under big data such as utility bills, a credit evaluation index system for small and micro enterprises has been constructed. Using ELECTRE-IN to eliminate the amplification effect of the preference index in ELECTRE-Ⅲ, and to perfect the process of constructing higher levels than relationships. Aiming at the problem of ELECTRE-IN’s strong subjectivity and complex ranking, firstly, the BWM-GiNi combination weighting method is used to determine the comprehensive weight of the evaluation index, and the threshold combination is determined according to the sample data range to enhance its objectivity. Secondly, the credibility method is used to replace the distillation algorithm to simplify the sorting process. Finally, a numerical example is used to sort a batch of small and micro enterprises, and the results of the numerical example verify the effectiveness of the model.

关 键 词：信用评价 小微企业 组合赋权 ELECTRE—IN 大数据 

分 类 号：F27[经济管理—企业管理]