线性时变参数DGM(1,1)变权组合预测模型  

Variable Weight Combination Prediction Model of Linear Time-Varying Parameters DGM(1,1)

作　　者：王丰效 

机构地区：[1]喀什大学数学与统计学院,新疆喀什844000

出　　处：《统计与决策》2017年第13期79-81,共3页Statistics & Decision

基　　金：国家社会科学基金资助项目(11XTJ001)

摘　　要：文章为了提高统计组合预测的拟合和预测精度,根据线性时变参数离散灰色预测模型的初值优化方法,给出了几个线性时变参数DGM(1,1)模型作为单项预测模型,进一步利用这些单项预测模型建立了一类变权线性时变参数组合预测方法。最后,将变权重线性时变参数组合预测方法应用于新疆生产建设兵团城镇化发展水平的组合预测,实例结果表明变权重线性时变参数组合预测方法具有较高的拟合精度。In order to improve the forecasting accuracy and the fitting of statistical combination prediction model, and accord- ing to the initial value optimization method of linear time-varying parameter discrete grey model （DGM）, this paper offers several linear time-varying parameters DGM （1,1） models as the single-item prediction model, and furthermore establishes a combined forecasting method of variable weight linear time-varying parameters by using these single-item prediction models. Finally, the pa- per applies this variable weight combination forecast method to prediction of urbanization development level of Xinjiang Construc- tion Corps. Living example analysis indicates that variable weight combination prediction method of linear time-varying parameters has a relatively higher fitting precision.

关 键 词：变权组合预测 线性时变DGM(1 1)模型 城镇化发展水平 

分 类 号：O212.1[理学—概率论与数理统计]