灰色绝对关联度组合预测模型的性质研究

组合预测能够充分利用各种不同的预测方法所提供的信息,以适当加权的形式得出组合模型,同时由于组合预测模型更关心某种预测信息的权重,从而使得预测问题所利用的信息更充分,更完备,模拟精度也更高。将灰色绝对关联分析应用到组合预测中,提出了基于灰色绝对关联度的组合预测模型,并在此基础上给出了灰色绝对关联优性组合预测、预测方法优超、冗余度等概念,给出了灰色绝对关联组合预测是非劣性组合预测及优性组合预测的一个充分条件,得出了冗余预测的判定定理。算例分析表明该方法是合理有效的。

Research on properties of combination forecasting model based on absolute of grey incidence

Combination forecasting can easily utilize information provided by all kinds of forecasting methods, and the combination forecasting model is derived by appropriate weights. Due to the fact that the weight of the forecasting information is concerned by the combination forecasting model, the information utilized in the process of forecasting is more adequate and more complete, which makes the simulation precision higher. It firstly applied abolute analysis of incidence into combination forecasting. Combination forecasting model based on absolute degree of incidence is proposed, some concepts such as superior combination forecasting of absolute analysis of incidence, dominant forecasting method of absolute analysis of incidence, redundant degree of absolute analysis of incidence, which is on the basis of combination forecasting model based on absolute degree of incidence is proposed, and some sufficient conditions of non-inferior and superior combination forecasting are given, some theorems which determine whether a method is a redundant forecasting are proved. Finally, it is illustrated by a numerical example that the method proposed is effective.