基于PLS和LOGISTIC回归模型的中小企业融资的信用风险度量  

Credit Risk Measurement of SME Financing Based on PLS and LOGISTIC Regression Model

作　　者：吴会咏[1,2] 刘艳春[1] 

机构地区：[1]辽宁大学商学院,辽宁沈阳110036 [2]沈阳化工大学数理系,辽宁沈阳110142

出　　处：《征信》2013年第7期7-11,共5页Credit Reference

基　　金：辽宁省高等教育学会"十二五"科研课题;区域产业结构调整实证研究(GHYB110003)

摘　　要：为了有效地度量中小企业融资的信用风险,选取代表中小企业的经营与发展能力、利润构成与盈利能力、资产与负债和现金流量等4个方面13个指标作为解释变量。由于解释变量之间存在较高的多重共线性和样本量偏小等情况,应用偏最小二乘法提取PLS成分,排除系统中的噪声干扰,构建一个度量信用风险的二分类因变量的Logistic模型。实际数据预测结果表明,该模型不仅具有良好的平稳性和准确性,而且具有较强的解释过程变化的能力。In order to effectively measure the credit risk of small and medium - sized enterprise （SME） financing, this paper selects such four aspects and 13 indicators as explanatory variables： SME management and development ability, profit composition and profitability, assets and liabilities and cash flow. Because there are stronger multi - collinearity and less samples in explanatory variables, by using PLS （ Partial Least Squares） for extracting PLS com- ponents, and eliminating the noise interference in the system, the paper constructs a sub-classification dependent vi- able Logistic Model to measure the credit risk. The prediction results of the actual data show that the model not only has good stability and accuracy but also has strong ability to explain the process of the change.

关 键 词：偏最小二乘 LOGISTIC回归 信用风险 中小企业融资 

分 类 号：F832.4[经济管理—金融学] F276.3