基于经验模态分解的航空公司旅客流量分析

Analysis of Airline Passenger Flow Based on Empirical Mode Decomposition

航空公司的旅客流量受宏观经济、季节性、竞争等诸多因素的影响,具有显著的非线性和非平稳性。不同于以往研究,本文提出应用经验模态分解(简称EMD)方法对航空公司的旅客流量进行分析。文中航空公司的旅客流量序列被分解为受宏观经济驱动的趋势项、受季节性因素影响的低频部分和受竞争等不确定性因素影响的高频部分。研究表明,EMD方法能有效地分解旅客流量序列,而且相比传统的线性预测模型,经过EMD分解的预测方法能得到更高的预测精度。

The passenger flow of an airline is influenced by many factors such as macro economics,seasonality,competition and so on.Unlike any traditional research on this area,this paper has proposed a new approach for the airline passenger flow analysis based on Empirical Mode Decomposition.The passenger flow of a major airline in China has been decomposed into three parts:trend driven by macro economics,low frequency part caused by seasonality and high frequency part caused by short-term uncertainty.The result shows EMD has been proven as an effective method to analyze the characteristics of the passenger flow in different time scales.