不同优化准则统计组合预测权系数优化  

Weight Coefficient Optimization of Combination Forecast by Different Criteria

作　　者：王丰效[1] 

机构地区：[1]喀什师范学院数学系,新疆喀什844000

出　　处：《数学的实践与认识》2013年第13期135-139,共5页Mathematics in Practice and Theory

基　　金：国家社会科学基金(11XTJ001)

摘　　要：不同的优化准则下,组合预测模型的权系数一般不同.为了兼顾不同准则的优点,根据均方误差,平均绝对误差以及平均绝对相对误差准则分别确定了组合预测模型的权系数向量,再利用这些权系数和层次分析方法构造新的组合预测模型的权系数.最后,通过几个应用实例,说明了这种建立组合预测模型方法的可行性和有效性.The weight coefficients are not same according to different optimization criteria. In order to take into account the advantages of different criteria, the weight coefficient vector of combination forecasting are obtained by the mean square error, the mean absolute error and the mean absolute percent error, respectively. Then the new weight coefficient of combination forecast model was constructed by these weights and AHP method. The application examples are given to show the feasibility and availability of this new combination forecast method.

关 键 词：组合预测 优化准则 权系数 

分 类 号：F224[经济管理—国民经济] F201