基于IOWA算子的税收收入的组合预测

针对1981~2012年我国税收收入的相关数据,采用回归模型、指数平滑预测和ARIMA模型三种单项预测方法进行建模预测.并结合组合预测理论,采用基于IOWA算子的组合预测模型进行预测.结果表明,基于诱导有序加权算术平均算子的组合预测模型的预测精度明显高于三种单项预测方法,说明了该方法用于税收收入预测的可行性和有效性,并采用5种有效性评价指标,检验了组合预测模型的预测效果.既然该方法可以通过调整组合权重提高预测精度,因此,在此基础上对今后3年的税收收入作出预测,发现今后3年税收仍会分别以15.59%、16.89%和16.77%的增长率增长.

Combination Forecasting on Tax Revenues According to IOWA Operator

For the relevant tax revenue data from 1981 to 2012 in China , this paper uses the regression model , exponential smoothing forecast and ARIMA model these three single forecasting methods to predict . Combined with the theory of combination predictions , this paper uses a combination of predictive models to predict based on IOWA operator , and the results indicate that induced ordered weighted arithmetic average operator combination forecasting model prediction accuracy is significantly higher than the three individual prediction method . Thus , this paper uses the method described on the feasibility and effectiveness of tax revenue forecast and evaluates the effectiveness of using five kinds of indicators , and then examines the effect of combination forecasting model to predict . This method can improve the prediction accuracy by adjusting the weight of a combination of power , therefore , on this basis , this paper forecasts the next three years of tax revenue , and find that the next three years＆#39; revenue will continue to grow respectively with the growth rate of 15.59% , 16.89% and 16.77% .