基于面积型中心误差准则的模糊组合预测模型及应用

Fuzzy Combination Forecasting Model Based on the Criterion of Area Type Center-error and Its Application

在复杂的系统预测中,系统信息往往以模糊不确定信息来刻画,而不是实数或者区间数。本文建立了一种新的三角模糊数信息的组合预测模型。针对传统的三角模糊数端点误差准则的缺陷,提出了三角模糊数的面积型中心的概念,并给出其计算表达式。在此基础上,获得了组合预测值与各个单项预测方法预测值的中心关系式。基于面积型误差准则,构建了新的模糊组合预测模型,并证明了该模型至少是非劣性的。最后,举例说明所提出组合预测模型具有更强的有效性,并对模型参数进行了灵敏度分析。

In the complex system forecasting,system information is often depicted by fuzzy uncertainty information instead of real number or interval number.This paper constructs a new combined forecasting model with triangular fuzzy number information.Due to the defects of endpoint error criterion for traditional triangular fuzzy number,the concept of area centers of the triangular fuzzy number is put forward,and their calculation expressions are given.On this basis,we obtain the area centers relationship between the combination forecasting and each single forecasting.A new fuzzy combination forecast model is established based on the criterion of area type center-errors,and it is proved that the model is at least non-inferior.Finally,an example of the proposed combination forecasting model is illustrated to show its better effectiveness,and sensitivity analysis is given for the parameters of the model.