基于因子分析的Logistic模型在中小板公司财务风险预警中的应用研究

发布时间：2017-12-26 18:02

本文关键词：基于因子分析的Logistic模型在中小板公司财务风险预警中的应用研究　出处：《华东交通大学》2016年硕士论文　论文类型：学位论文

【摘要】：本文以中小板上市公司作为研究对象,选取了被首次警告退市的公司前3年的财务数据,结合中小板公司的财务特点,从盈利水平、现金流量状况、营运能力、成长能力、偿债能力这五个部分,选取了具有代表性、综合性和对中小企业的财务特征敏感的33个财务变量,运用logistic模型进行财务风险预警的应用研究。运用logistic方法分别对3年的财务数据进行建模分析,分别建立不同时期的logistic财务风险预警模型,并利用对应的预测组样本对模型进行了验证。结果表明,模型训练的正确率/预测的正确率分别为:100.00%/86.67%、86.67%/76.67%、83.33%/70.00%。针对logistic模型存在的多重共线性、应用泛化能力不足等缺陷,采用因子分析的方法对财务指标进行公因子提取,优化logistic建模过程,建立不同时期、基于因子分析的logistic财务风险预警模型。最终得到模型的训练正确率/预测正确率分别为:100.00%/86.67%、100.00%/90.00%、86.67%/73.33%。总体来说,基于因子分析的logistic财务风险预警模型的预测精度比原模型的要高。同时,将三年的预测判别结果进行综合考虑,比较模型的综合判别能力。以不同时期模型的预测正确率作为权重,将不同的风险预警模型的判断结果进行综合,构建综合判断方程。验证结果表明,logistic模型的预测结果综合判断方程对全部样本公司的预测正确率为81.67%,而基于因子分析的logistic财务风险预警模型的综合判断方程对样本公司的预测正确率为86.67%,基于因子分析的logistic模型综合结果判断预测精度比原logistic模型的要高。最后,对本文的工作进行了总结,对模型的应用与研究的不足、下一步的研究等问题进行了说明。
[Abstract]:In this paper, the SME board listed companies as the research object, selected financial data is the first warning of delisting of the company 3 years ago, combined with the financial characteristics of small companies, from the five part of profitability, cash flow, operation ability, growth ability, solvency, representative, comprehensive and financial the characteristics of the small and medium-sized enterprise sensitive 33 financial variables, application research based on logistic model of financial risk early warning. Logistic method is used to model and analyze the 3 years' financial data, and logistic financial risk early warning models are established in different periods. The corresponding prediction group samples are used to validate the model. The results show that the correct rate of model training is 100.00%/86.67%, 86.67%/76.67% and 83.33%/70.00% respectively. Aiming at the defects of multiple collinearity and application generalization ability of logistic model, we use factor analysis method to extract common factors of financial indicators, optimize logistic modeling process, and establish logistic financial risk early warning model based on factor analysis in different periods. The correct rate of training and prediction of the model is 100.00%/86.67%, 100.00%/90.00% and 86.67%/73.33% respectively. In general, the prediction accuracy of logistic financial risk early warning model based on factor analysis is higher than that of the original model. At the same time, the three year prediction results are taken into consideration, and the comprehensive discriminant ability of the model is compared. The prediction accuracy of different period models is taken as the weight, and the judgment results of different risk early warning models are synthesized and the comprehensive judgment equation is constructed. The verification results show that the prediction results of logistic model prediction accuracy rate equation for all Sample Firms comprehensive judgment was 81.67%, and the prediction of logistic comprehensive judgment of financial risk early warning model of factor analysis of the equation of the Sample Firms based on the correct rate is 86.67%, determine the prediction accuracy than the original logistic model to logistic model results based on factor analysis. Finally, the paper summarizes the work of this paper, and explains the problems of the application and research of the model, the next step of the research.
【学位授予单位】：华东交通大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：F275

【相似文献】