基于奇异谱分析的我国航空客运量集成预测模型

发布时间：2018-09-04 10:56

【摘要】：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.
[Abstract]:Aiming at the problems of time series including noise and the instability of single model, this paper presents an integrated prediction model based on singular spectrum analysis (SSA), and applies it to the prediction of annual air passenger volume in China. Firstly, SSA method is used to decompose and reconstruct the original time series, and a noisy time series is obtained. Then it is used as the input of single model such as single integral autoregressive moving average model, (ARIMA), support vector regression model, (SVR) Holt-Winters method, (HW) and so on. Then the weighted average ensemble prediction method (WA) is used to integrate the prediction results of three single models. By comparing with the single model, the model based on empirical mode decomposition method (EMD) and the simple average integrated prediction method (SA), it is found that the proposed model has high prediction accuracy and stable prediction performance. Finally, the model is used to forecast the annual air passenger volume in 2014-2016.
【作者单位】：上海大学管理学院;中国科学院大学经济与管理学院;中国科学院数学与系统科学研究院;
【基金】：国家自然科学基金(71422015,71373262) 国家数学与交叉科学研究中心全球宏观经济监测预测与政策模拟平台项目~~
【分类号】：F562

【相似文献】