税收组合预测仿真研究

Combination Prediction Simulation Research on Tax

研究税收组合准确预测问题,税收受到经济、政策等多种因素作用,单一模型只能反映税收变化的部分信息,预测精度低。为提高税收预测精度,提出一种灰色模型和多元线性回归的税收组合预测模型。首先分别采用灰色模型和多元线性回归对税收进行预测,然后采用拟合误差的方差最小原则确定两种模型的权重系数,最后得到组合模型的税收预测值。采用我国1988～2010年税收数据进行仿真,结果表明,相对于单一预测模型,组合预测模型不仅提高了税收的预测精度,而且为全面税收的复杂、动态准确预测提供了依据。

Research on combination forecasting problem of tax.In order to improve the prediction precision of tax,this paper proposed a tax combination prediction model by grey model and multiple linear regression.Firstly,grey model and multiple linear regression were used to predict the tax,then the mode weights were determined by fitting error of minimum variance principle finally get tax prediction value.The model was tested with 1988~2010 tax data,and the simulation results show that,compared with single forecast model,the proposed model improves the tax prediction precision and can comprehensively reflect the complex,dynamic change rule of tax.