本文运用多元统计分析方法,通过对所选取的中小上市公司数据进行聚类分析和判别分析,建立违约模型。实证分析得出的判别函数较为简洁直观,能对新的企业数据进行违约特征的判别,并可知此种判别方法的判别误差及错判概率。 In this paper, multivariate statistical analysis method, through the selected small and medium-sized listed companies data clustering analysis and discriminant analysis, the establishment of default model. The discriminant functions derived from the empirical analysis are relatively simple and intuitive, and can discriminate the default characteristics of the new enterprise data. The discriminant errors and the probability of misjudgment of such discriminant methods can also be known.