论文部分内容阅读
结合模糊软集合理论建立税收收入的组合预测模型,根据税收收入的特点,代表性地选择了Elman回归神经网络模型、含政策虚拟变量的自回归模型、ARIMA(1,1,1)的时间序列模型、多因素SVM回归模型这四种模型作为组合预测中的单一模型,并以1980年到2008年的税收收入等相关数据为背景进行了说明和分析.结果表明该组合预测模型能有效减小预测误差,为税收工作实践提供了一个应用研究工具,并推广和丰富了软集合理论在税收经济模型研制中的实际应用.
According to the characteristics of tax revenue, Elman regression neural network model, autoregressive model with policy dummy variable, time series of ARIMA (1,1,1) are selected according to the combination of fuzzy soft set theory and tax revenue. Model and multivariate SVM regression model as the single model in combined forecasting and the related data of tax revenue from 1980 to 2008. The results show that the combined forecasting model can effectively reduce Forecast error, which provides an applied research tool for tax practice and popularizes and enriches the practical application of soft set theory in the development of tax economic model.