摘要
为了克服传统财务危机预警模型在假设前提、样本容量、泛化能力等方面的缺陷,应用非线性SVM构建了财务危机预警模型。该模型以偿债能力、营运能力、盈利能力、现金能力和成长能力等五方面的15个财务指标作为输入变量,以上市公司是否被特别处理(ST)作为输出变量,实证分析表明:该模型具有100%的训练精度和90%的验证精度,学习和预测能力良好。
To overcome the shortages of traditional financial prediction models such as strict hypothesis, big sample size and poor generalization ability, this paper establishes a new model for listed company using non - linear support vector machine （SVM）. The input variables of this model include 15 financial indexes in 5 dimensions as follows： debt - paying, operation, profit - making, cash and growth; the output variable is defined as whether the listed company gets special treatment （ST） or not. The positive analysis shows the training and validation accuracies of the model are 100% and 90% respectively, which concludes that learning and generalization abilities of this model are excellent.
出处
《统计与信息论坛》
CSSCI
2009年第6期49-53,共5页
Journal of Statistics and Information
基金
黑龙江省科技计划项目<哈尔滨大庆国家高新区管理体系研究>(GC06D209)
