基于改进的组合预测误差区间数组合预测模型

Interval Number Combination Forecast Model Based on Improved Combined Forecasting Error

文章通过引入偏好系数,将区间数中点及半径的预测误差进行组合,构造出区间数的组合预测误差,将组合的预测误差带入灰关联度公式,得出两组区间数的灰关联度。并建立了基于灰色关联度的定权区间组合预测模型以及基于广义诱导有序加权平均算子和灰色关联度的变权区间数组合预测模型。不仅从理论上证明模型的有效性,而且通过实例验证了模型的有效性。

By introducing preference coefficient this paper combines the prediction error of midpoint of interval number as well as radius to construct combination forecast error of interval number, and plugs the combined prediction error into grey rela- tional formula and gets the grey correlation degree of the two groups of interval. The paper also builds weighting interval combina- tion forecast model based on grey correlation degree and variable weight interval combination forecast model based on GIOWA op- erator and grey correlation degree. The validity of the proposed model is not only theoretically proved, but its effectiveness is also verified through empirical examples.