CAJ | 학술논문

中小企业集合债券融资个体的信用风险度量面临小样本、非线性、高维数等现实问题,传统的评估方法很难适用.为了弥补传统评估方法的不足,提高信用风险度量的准确性,建立了适用性更强的信用风险评估指标体系,并引入基于统计学习理论的 SVM 模型对融资个体信用风险进行度量.选取径向基核函数作为支持向量机的核函数,通过数据的转化与缩放、参数的优选,最终获得了分类效果比较好的中小企业集合债券融资个体信用风险度量模型.经实际数据检验,模型的预测准确率为90.77%,具有较强的适用性.
중소기업집합채권융자개체적신용풍험도량면림소양본、비선성、고유수등현실문제,전통적평고방법흔난괄용.위료미보전통평고방법적불족,제고신용풍험도량적준학성,건립료괄용성경강적신용풍험평고지표체계,병인입기우통계학습이론적 SVM 모형대융자개체신용풍험진행도량.선취경향기핵함수작위지지향량궤적핵함수,통과수거적전화여축방、삼수적우선,최종획득료분류효과비교호적중소기업집합채권융자개체신용풍험도량모형.경실제수거검험,모형적예측준학솔위90.77%,구유교강적괄용성.
@@@@ The small and median enterprises assemble bond financing of individual credit risk facing many realistic problems, such as small samples, nonlinear, high dimensions and so on, to which the traditional evaluation method is difficult to apply. In order to make up for the shortcomings of the traditional evaluation methods, and enhance the credit risk measurement accuracy, this paper established a more applicable credit risk evaluation index system, and introduced support vector machine model based on statistical learning theory. Having selected the radial basis kernel function as the kernel function of support vector machine, through the data conversion and scaling, parameter optimization, the authors finally obtained good classification effect of small and median enterprises assemble bond financing of individual credit risk measurement model. After the test of actual data, the forecasting accuracy reached 90.77%, the model has strong applicability.