基于Prophet的民航商务旅客出行量预测研究  

Research on Prediction of Civil Aviation Business Passenger Travel Volume Based on Prophet

作　　者：鲍斌 云雄[1] 甘国操 谢佳 刘辉 BAO Bin;YUN Xiong;GAN Guo-cao;XIE Jia;LIU Hui(Chongqing R&D Center of TravelSky Technology Limited,Chongqing 401000,China;TravelSky Technology Limited,Beijing 101000,China;Key Laboratory of Intelligent Application Technologyfor Civil Aviation Passenger Services,Civil Aviation Administration of China,Beijing 101000,China)

机构地区：[1]中国民航信息网络股份有限公司重庆研发中心,重庆401000 [2]中国民航信息网络股份有限公司,北京101000 [3]中国民用航空局民航旅客服务智能化应用技术重点实验室,北京101000

出　　处：《航空计算技术》2024年第2期79-82,87,共5页Aeronautical Computing Technique

基　　金：民航重点科技项目资助(P3DA043)。

摘　　要：随着民航业竞争日益加剧,民航市场需求的细分和预测对于民航业的发展具有重要意义。而商务旅客作为民航客源重要组成部分,民航商务旅客出行量的准确预测,为民航业的科学规划、运营管理和收益管理等提供重要的决策依据。为探索一种准确率高、可解释性强的预测算法,对民航商务旅客出行量数据进行了可视化和分析,结合数据特点提出了基于Prophet模型的民航商务旅客出行量预测方法,并通过真实的民航商务旅客出行量数据进行实验和分析。实验结果表明,基于Prophet的模型与同类的ARIMA模型和LSTM模型相比,具有更高的预测精准度,同时具有更强的可解释性,证明了方法的有效性。With the intensifying competition in the civil aviation industry,the segmentation and prediction of market demand for civil aviation becomes crucial for its development.As a significant component of civil aviation passenger source,accurately predicting the travel volume of business passengers provides a vital foundation for scientific planning,operational management,and revenue management within the industry.In order to explore a highly accurate and interpretable prediction algorithm,this study visualizes and analyzes data on travel volume of business passengers in civil aviation.It proposes a prediction method based on the Prophet model that incorporates data characteristics and conducts experiments and analyses using real-world travel volume data from business passengers in civil aviation.The experimental results demonstrate that the Prophet-based model exhibits higher prediction accuracy and stronger interpretability compared to similar models such as ARIMA and LSTM,which validate the effectiveness of the method.

关 键 词：商务旅客 出行量预测 时间序列模型 Prophet模型 

分 类 号：V354[航空宇航科学与技术—人机与环境工程]