基于灰色关联度的组合预测模型的性质

基于灰色关联度的组合预测模型是研究组合预测方法的一个新思路,实例结果表明它是一种有效的组合预测方法.针对该模型,提出新的优性组合预测、预测方法优超、冗余度等概念;在一定的条件下, 证明模型的任一个可行解对应的组合预测至少是非劣性组合预测; 给出优性组合预测存在的一个充分条件, 最后证明冗余预测方法的一个判定定理,这对组合预测方法有效信息的提取具有重要的意义.

Properties of combination forecasting model based on degree of grey incidence

It is a new idea to study combination forecasting based on degree of grey incidence. Numerical examples have shown that it is a kind of effective combination forecasting method. In this paper, some new concepts are proposed, such as superior combination forecasting, dominant forecasting method, redundant degree, etc. Under certain conditions, it is proved that combination forecasting corresponding to the arbitrary feasible solution of this model is at least non inferior. The sufficient conditions of existence of superior combination forecasting are also given. Finally the determining theorem of redundant forecasting method is proved, which is significant in extracting effective information in combination forecasting.