基于免疫粒子群算法的组合预测模型研究  

Research of combination forecasting model based on IA-PSO

作　　者：赵明[1] 黄启来[1] 张作刚[2] 

机构地区：[1]海军航空工程学院,山东烟台264001 [2]海军航空工程学院青岛校区,山东青岛266041

出　　处：《电子设计工程》2013年第3期15-16,19,共3页Electronic Design Engineering

摘　　要：使用一种预测模型对一问题进行预测时,往往只是考虑了该问题所具有的许多侧面中某一个或某几个。使得一种预测模型刻画研究对象的未来发展规律往往会存在着一定的局限性。所以在已经研究出得在众多预测模型中,分清某一预测模型对于揭示研究对象的变化规律所存在的不同优势,然后根据预测问题的实际需要将一些不同的预测模型进行一定机制下的有机组合,形成组合预测模型。文中利用免疫粒子群算法来计算组合预测中各预测模型的权值。可以很好地提高预测精度,更加贴切描述预测问题的演化规律。One prediction model for the problem of forecasting, only considering one or Several sides of the problem. Make a prediction model to characterize the research object of the future development of the law often exist certain limitations. So have developed in numerous forecast model, distinguishing between a prediction model for revealing the research object changes the existence of distinct advantage.Then according to the prediction problem actually required to be a number of different prediction models under certain mechanisms of organic combination, format the combination forecasting model. This paper uses IA-PSO to calculate the weight in the forecasting model. It can improve the prediction accuracy and more appropriate to describe the evolution of the prediction problem.

关 键 词：免疫算法 粒子群 组合预测 模型 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]