基于数据挖掘的航班延误预警管理分析

Flight Delays Early Warning Management and Analysis Based on Data Mining

目前国内对航空公司航班延误状态的描述缺乏统一的计算模型,对航班延误结果的评价也不是很明确。基于航空公司大量历史运行数据,结合数据挖掘中的预测模型建立方法,采用层次分析法(Analytic Hierarchy Proces,AHP)提出了以延误率、平均延误时间、延误旅客人数为评价指标的航班延误状态描述方法,并运用马尔可夫(Markov)链对评价指标进行预测。利用模糊层次分析法(Fuzzy Analytic Hierarchy Process,FAHP)得到各指标权重,结合模糊综合评判法对航班延误状况进行综合评价,建立航班延误预警指标体系。仿真实验与结果分析表明:预警指标能较准确地反映航班延误状况,评价结果客观,可为航空公司航班延误预警管理提供理论与方法支持。

It is still a hard task to model domestic flight delay and to evaluate flight consequence in an authorized way at present.In this paper,according to a large number of airlines' s historical operation data,a mathematical flight delay forecast model was derived to estimate the delay of each flight by analyzing the main factors influencing the flight delay,the consideration of the scheduled flight delay rate,the average time of delay and the delay passengers,and the Markov theory which is based on data mining prediction model establishment method and is used to forecast the evaluation index.The weights are gained by using fuzzy analytical hierarchy process(FAHP),and the condition of flight delay is evaluated based on fuzzy comprehensive evaluation.Finally,the flight delay alarming index system was constructed.The experiment of the developed delay forecast model shows that these alarming indexes can reflect the condition of flight delay accurately,and the outcome is objective,so it can provide support for flight delay early warning management theory and method.