基于DTW的机场延误特性分析

摘要

针对国内航班延误现状,尝试从机场延误时间序列数据本身分析,挖掘机场延误时间序列特性,找出与延误时间相关的影响因素,重点考察机场离港航班整体延误情况.基于动态时间规整(DTW)方法,分析机场延误时间序列的整体相似特性;在时序相似性度量的基础上,结合皮尔逊相关系数探究在延误情况相似情况下与机场延误时间序列相关的其他时序因素.结果表明,天气因素对于机场延误时间影响较大,且延误总时间相近的月份之间,机场时间序列也更为相似;实际进离港架次和前序延误航班对于机场延误时间序列影响较大,而发生机上延误及过站冗余时间的关联性较小.%For analyzing the the nature of airport delay time series itself ,finding factors related to the de-parture delay,Dynamic Time Warping (DTW) distance is considered as a desirable choice in exhibiting similar delay patterns occurring at different time periods.In this paper,Pearson correlation coefficient is used on the basis of the time series similarity measure.As a result,the potentially underlying factors con-nected to the departure delay might be identified.The results show that weather factors have a great influ-ence on the departure delay and the DTW distance of airport delay time series sharing similar delay situa-tions should be more close.Besides,the actual airport throughout and propagated delays greatly affect de-parture delays,but the delays on board and the extra transit time have little impact.