基于预测误差度的IGOWLA算子组合预测模型

IGOWLA Operator Combined Forecasting Model Based on Predictive Error Degree

针对组合预测的误差不一定比其单项方法的误差小这一问题,基于"预测有效度",提出预测误差度"的概念;将诱导广义有序加权对数平均(IGOWLA)算子与预测误差度结合,建立了新的最优组合预测模型,实证结果表明该模型的误差比其单项方法的误差更小且更稳定.

The error of combination forecasting is not necessarily smaller than the error of its single method. Based on the prediction validity, the concept of forecasting error degree is put forward, and a new optimal combination forecasting model is established by combining the induced generalized ordered weighted logarithm averaging （IGOWLA） operator with the predictive error degree. The example results show that the model error is smaller and more stable than the single method.