基于因子分析的Logistic违约概率模型  

Logistic Default Probability Model Based on Factor Analysis

作　　者：张颖[1] 马玉林[2] 

机构地区：[1]同济大学经济与管理学院,上海200092 [2]山东财政学院统计与数理学院,济南250014

出　　处：《桂林工学院学报》2010年第1期174-178,共5页Journal of Guilin University of Technology

基　　金：山东省软科学研究计划资助项目(2008RKB242)

摘　　要：选取我国泛北部湾6省(区)360家上市公司的财务数据,采用能反映公司信用特征的52个财务指标作为原始变量,通过因子分析,在提取6个主成分基础上,利用Logistic模型建立了商业银行上市公司贷款客户违约概率函数,经检验该信用风险模型的总体正确率为90.2%,对不违约的预测正确率为94.1%,对违约的预测正确率为86.3%。Fifty-two financial indicators reflecting company credit characteristics are used as original variables to formulate a logistic model to measure default probability of listed companies in which six principal components are extracted by factor analysis.Using financial data of listed companies in six provinces of Pan Beibu Gulf economic cooperation area,the empirical test results show that the logistic model has very credible ability to identify,measure and predict credit risks,with accuracy of 90.2%.

关 键 词：违约概率 因子分析 LOGISTIC模型 

分 类 号：O29[理学—应用数学] F83[理学—数学]