论文部分内容阅读
本文提出一种多模型自适应预报方法,预报系统由几个并行的多步预报器组成,最终预报由Bayes决策律决定。本法适用于随机快时变参数动态过程的预报,在我国工业用电量长期预报中应用此法获得了良好的效果。
In this paper, a multi-model adaptive forecasting method is proposed. The forecasting system consists of several parallel multi-step predictors. The final forecast is determined by Bayes decision-making law. This method is suitable for forecasting the dynamic process of stochastic fast time-varying parameters. The method has been applied to the long-term forecast of industrial electricity consumption in China.