信贷风险评价指标权重的聚类分析  

Cluster Analysis for Weight of Credit Risk Evaluation Index

作　　者：迟国泰[1] 郝君[1] 徐琤[1] 朱战宇[1] 

机构地区：[1]大连理工大学管理学院,大连116024

出　　处：《系统工程理论方法应用》2001年第1期64-67,共4页Systems Engineering Theory·Methodology·Applications

基　　金：国家自然科学基金专项基金资助项目 !(70 0 42 0 0 2和79942 0 0 9) ;加拿大国际开发署 (CIDA)中 -加大学与产业合作项目 (CCUIPP)

摘　　要：信贷风险评价是银行重要的管理活动 ,评价指标的权重评估又是决定评价结果是否科学的关键环节。本文在分析同类研究弊端的基础上 ,利用相似系数矩阵对众多专家评价指标的权重意见进行筛选 ,解决了信贷风险评价体系指标权重的合理分配问题。其特色在于 :一是对专家意见进行聚类分析 ,而不是像通常那样不加处理地简单平均 ;二是在剔除某些专家权重意见时 ,结合实际问题的具体要求 ,在聚类分析的基础上 ,提出了一种更为简化的方法 ,易于在实践中推广 ;三是对于通过任何方法得出的群体权重意见 。Evaluation of credit risk is an important management activity of the bank. Weight of evaluation index is the key link in determining if the evaluation result is scientific. Based on analyzing the disadvantages of relative research, this paper selects the group specialists' advice of evaluation index with a matrix of similarity factor, and solves the problem of reasonable distribution of index weight in credit risk evaluation system. Characteristics of this paper lies in:First, it does not average the specialists' advice without handling, but analyzes it with principle of cluster. Second, in case of getting rid of a few specialists' advice, one more simplified method is put forward according to the specific requirement. Third, this method can also be the subsequent handling procedure of weight advice no matter how it is got.

关 键 词：信贷风险 风险评价 聚类分析 银行 权重 

分 类 号：F830.5[经济管理—金融学]