摘要
本文依据6种财务可持续成长模型提炼预警指标,基于2003～2007年我国A股市场新增ST公司样本及配对样本的前3年数据,对各模型的驱动因素进行综合预处理,分别得到ST前1～3年的预警指标体系。然后,建立预警模型进行回判和外推。结果表明,各种分析方法在ST前3年可获得较高的判断准确性;财务危机与可持续成长驱动指标之间存在着高度的非线性关系;可持续成长模型能够为开发标准化预警指标体系提供依据。
This paper introduces six sustainable growth models to abstract variables. After synthetically pretreating variables to get the index systems before one to three years, this paper uses four dominating methods of multivariate discrimination, logistic regression, neural network and least squares support vector machine to carry out empirical study for bankruptcy prediction. Result shows that these four methods have good classification abilities before three years and there exist great nonlinear relations between finance distress and driving indexes. Sustainable growth model can be used to exploit standard index systems for early warning.
出处
《经济管理》
CSSCI
北大核心
2008年第19期83-88,共6页
Business and Management Journal ( BMJ )
基金
教育部博士学科点专项科研基金项目“动态竞争环境下基于核心能力战略的企业集团危机预警系统研究”(20050286038)
关键词
可持续成长模型
多元判别
逻辑回归
神经网络
最小二乘支持向量机
sustainable growth model
multivariate discrimination
logistic regression
neural network
least squares support vector machine
