摘要
为了克服传统的单项预测方法选取固定参数时的不足,在广义诱导有序加权对数平均算子(IGOWLA算子)的基础上,引入贴近度以及λ次幂误差,构建了基于一种贴近度的IGOWLA算子的最优组合预测模型,并给出了该模型的预测精度、优性及非劣性定义.实例分析表明,该组合预测模型优于传统的单项预测模型,能够充分利用各个单项预测方法的信息并提高预测精度,是一种优性组合预测.
In order to overcome the single forecasting method to take the fixed weight coefficient of defects, a new combination forecasting model was proposed based on one kind of closeness degreeand induced geometric ordered weighted logarithm averaging （IGOWLA） operator by combining IGOWLA operator, power error and the degree of closeness. At the same time, the prediction aceuracy of the model, the conception of superior combination forecasting and noinferior combination forecastingwas given. Finally, an example is illustrated by using the model. The example showed that the combination forecast model can make full use of every single forecasting method of informationand the forecasting precision is superior to the traditional single forecasting model, therefore, it was superior combination forecasting.
出处
《延边大学学报（自然科学版）》
CAS
2017年第1期19-24,共6页
Journal of Yanbian University（Natural Science Edition）
基金
国家社会科学基金资助项目(12BTJ008)
关键词
贴近度
组合预测
IGOWLA算子
closeness degree
combination forecasting
IGOWLA operator
