摘要
在递归等权组合预测方法(REW法) ̄[1]和递归方差倒数组合预测方法(RVRW法) ̄[2]的思路基础上,以方差的幂函数倒数构造组合权重,进一步提出了广义递归方差倒数组合预测方法(GeneralizedRecursiveVarianceReciprocalWeighting,即GRVRW法),给出了有关的迭代计算方法。该方法更一般地体现了以预测精度作为组合权重依据的思想,将REW法和RVRW法概括为其特殊情况,实例表明,GRVRW法可以明显地提高预测精度。
Based on basic idea of Recursive EquaI Weighting(REW) method and Recursive VarianceReciprocal Weighting(RVRW) method,the Generalized RVRW(GRVRW) method is developed byintroducing the idea of variance power reciprocal weighting,and.related iterative formulas are proposed.Thestudy shows that both REW method and RVRW method are the special cases of GRVRW method.Finally,an illustrative example is given to show the effectiveness of this method.
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
1995年第2期211-217,共7页
Journal of University of Electronic Science and Technology of China
基金
国家自然科学基金
关键词
递归等权组合
预测
方法
递归方差
倒数组合
combination forecasting
iterative algorithm
REW method
RVRW method
GRVRW method