论文部分内容阅读
针对时间序列包含噪声以及单一模型可能存在预测表现不稳定的问题,本文提出了一个基于奇异谱分析(SSA)的集成预测模型,并将其运用于我国年度航空客运量的预测中.首先,采用SSA方法对原始时间序列进行分解和重构,得到一个剔除噪声的时间序列,然后将其作为单整自回归移动平均模型(ARIMA)、支持向量回归模型(SVR)、Holt-Winters方法(HW)等单一模型的输入并进行预测,接着再采用加权平均集成预测方法(WA)将三种单一模型的预测结果进行综合集成.通过与各单一模型、基于经验模态分解方法(EMD)的模型以及简单平均集成预测方法(SA)的预测结果进行对比发现,本文所建模型具有较高的预测精度和较稳定的预测表现.最后,采用本文的模型对我国2014-2016年年度航空客运量进行了预测.
In order to solve the problem that the time series contains noise and the single model may have unstable prediction performance, an integrated prediction model based on singular spectrum analysis (SSA) is proposed and applied to the prediction of annual air passenger traffic in China.Firstly, SSA method is used to decompose and reconstruct the original time series to obtain a time series of noise elimination, which is then used as an ARIMA, SVR, Holt-Winters method (HW) (WA) is used to integrate the prediction results of three single models.According to the model of each single model, empirical mode decomposition (EMD) method and Compared with the prediction results of simple average integrated forecasting method (SA), the model proposed in this paper has higher prediction accuracy and more stable forecasting performance.Finally, using the model of this paper, the annual air passenger traffic volume of China from 2014 to 2016 prediction.