欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
以中国A股上市公司财务数据为研究样本,运用主成分分析提炼出对财务危机具有显著影响的指标变量作为输入变量,建立AdaBoost财务危机预测整体模型.最后将BPNN和GA-BPNN作为实验对比模型进行比较分析,进而运用分类准确率对三种公司财务危机预测模型的预测精度进行检验、评价.结果表明,三种财务危机预测模型对我国上市公司财务危机预测的分类准确率分别为92％、84％、88％;相比其他两种财务危机预测模型,Ada-Boost财务危机预测整体模型效果更好,说明AdaBoost算法能够提高单种财务危机预测模型的预测准确性,有利于解决公司财务危机问题.
英文摘要：
This paper attempts to construct an AdaBoost ensemble model to predict the financial distress. The Chinese listed companies are regarded as the initial sample. Then extract eight significance financial indicators from the optional indicators by principle component analysis, and regard the eight financial indicators as the AdaBoost ensemble model＇ s input variables. The classification accuracy is regarded as the evaluation criteria to test the AdaBoost ensemble model, and the forecasts obtained by the AdaBoost ensemble model have been compared with BPNN, GA-BPNN models. The result shows that each of three models＇ classification accuracy rates were84%, 92%, 88%, respectively. The computational results demonstrate that AdaBoost ensemble model can markedly improve the accuracy of single classifier, and it is more helpful to save the companies＇ financial distress problem.
同期刊论文项目
同项目期刊论文
期刊信息
位置：