摘要
本文研究了组合预测的基本理论,提出了有效的最优权重确定方法。基于贝叶斯理论分析,在相当一般的条件下,本文推导了非线性组合预测方式。该组合具有相当一般性并将传统算术加权平均、几何加权平均等组合预测方式作为特例。实例分析证实了本文方法的有效性。
In this paper, basic theory of canbined forecasts is investigated andapproach to determining optimal weights is presented. Bayesian analysis in this paperleads to nonlinear combination of forecasts under general conditions, which considerstraditional weighted arithmetic averaging formula and weighted geometric averagingformula as its special cases. Illustrative example proves its efficiency.
出处
《系统工程学报》
CSCD
1996年第1期28-35,共8页
Journal of Systems Engineering
基金
国家自然科学基金
关键词
组合预测
贝叶斯分析
非线性组合
经济预测
combined forecasts, Bayesian analysis, nonlinear combination, weight