中文核心期刊
CSCD来源期刊
中国科技核心期刊
RCCSE中国核心学术期刊

重庆交通大学学报(自然科学版) ›› 2021, Vol. 40 ›› Issue (12): 12-18.DOI: 10.3969/j.issn.1674-0696.2021.12.03

• 交通+大数据人工智能 • 上一篇    

基于KNN的机场航班短期延误风险预测

刘继新1,2,杨光1,2   

  1. (1. 南京航空航天大学 民航学院,江苏 南京 210016; 2. 国家空管飞行流量管理技术重点实验室,江苏 南京 210016)
  • 收稿日期:2020-06-23 修回日期:2020-10-27 发布日期:2021-12-27
  • 作者简介:刘继新(1966—),男,安徽滁州人,副教授,主要从事交通运输规划与管理方面的研究。E-mail:larryljx66@nuaa.edu.cn 通信作者:杨光(1994—),女,陕西汉中人,硕士,主要从事交通运输规划与管理方面的研究。E-mail:450983429@qq.com
  • 基金资助:
    南京航空航天大学研究生创新基地 (实验室) 开放基金项目 (kfjj20190707)

Short-Term Flight Delay Risk Forecast Based on KNN

LIU Jixin1, 2,YANG Guang1, 2   

  1. (1. College of Civil Aviation, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, Jiangsu, China; 2. National Key Lab in Air Traffic Flow Management, Nanjing 210016, Jiangsu, China)
  • Received:2020-06-23 Revised:2020-10-27 Published:2021-12-27

摘要: 针对当前民航延误问题仍然十分突出的现状,将航班发生短期延误的时长及其发生概率视为一种风险问题,考虑恶劣天气影响,将航班随外部运行环境的随动性变化与挖掘航班历史数据结合,对该风险进行评估预测。选用主成分分析法对众多因素进行分析和筛选,确定延误的关键影响因素,作为分类预测算法的样本属性。采用KNN算法建模,结合历史航班运行数据和天气数据,对机场短期内离港的单航班起飞延误状况及风险值进行预测。实验结果表明,主成分分析法能够在变量间存在信息重叠的情况下较为准确地找出关键因子,在此基础上采用KNN算法对航班延误风险进行预测,能够取得较好效果,具有实际应用意义。

关键词: 交通运输工程;航班延误;风险预测;主成分分析法;KNN

Abstract: In view of the current situation that civil aviation delay is still very prominent, the duration and probability of short-term flight delay are regarded as a risk problem. Considering the impact of bad weather, the risk was evaluated and predicted by combining the follow-up change of flight with external operation environment with mining flight history data. The principal component analysis method was used to analyze and screen many factors, and determine the key influencing factors of delay as the sample attributes of classification prediction algorithm. The KNN algorithm was used to build a model, which combined historical flight operation data and weather data to predict the departure delay and risk value of single flight leaving the airport in a short time. The experiment results show that the principal component analysis method can accurately find the key factors when there is information overlap between the variables. On this basis, KNN algorithm is used to predict the flight delay risk, which can achieve good results and has practical significance.

Key words: traffic and transportation engineering; flight delay; risk forecast; PCA (principal component analysis); KNN algorithm

中图分类号: