基于网络搜索引擎的大型活动客流规模预测  

Passenger flow prediction for large-scale events based on network search engine

作　　者：董春娇[1] 刘晓珂 常乃心 李林玉 DONG Chunjiao;LIU Xiaoke;CHANG Naixin;LI Linyu(Key Laboratory of Transport Industry of Big Data Application Technologies for Comprehensive Transport,Beijing Jiaotong University,Beijing 100044,China)

机构地区：[1]北京交通大学综合交通运输大数据应用技术交通运输行业重点实验室,北京100044

出　　处：《北京交通大学学报》2022年第4期52-59,共8页JOURNAL OF BEIJING JIAOTONG UNIVERSITY

基　　金：国家重点研发计划(2019YFF0301400)。

摘　　要：在对持续性大型活动进行交通需求预测时,会出现因缺失资料而导致预测困难的情况.针对该问题从综合搜索指数,潜在出行人口规模和吸引度3个方面,分析了持续性大型活动客流规模影响因素,根据多项式分布滞后(Polynomial Distributed Lag,PDL)模型构建了基于网络搜索引擎的大型活动客流规模预测模型.以2019年北京世界园艺博览会为例,利用北京市2012年至2019年北京客流量对预测模型进行参数标定,预测世园会5月客流规模.结果表明:网络搜索引擎数据与大型活动客流规模存在相关关系.对于无历史数据的持续性大型活动,该预测模型的平均绝对百分比误差(Mean Absolute Percentage Error,MAPE)在8%以内,有较高的预测水平,能够为有关部门提供决策依据.When predicting the traffic demand for continuous large-scale events, the prediction is usually difficult due to the lack of relevant data. To solve this problem, this paper analyzes three influencing factors for passenger flow in continuous large-scale events: comprehensive search index, potential travel population scale and attraction degree. Then, a Polynomial Distributed Lag(PDL) model based on network search engine is built to predict the passenger flow in large-scale events. Taking the International Horticultural Exhibition 2019 in Beijing as an example, the prediction model is calibrated with data of the Beijing passenger flow from 2012 to 2019, thus predicting the passenger flow of the Expo in May. The results show that there is a correlation between the network search engine data and the scale of large-scale event passenger flow. For continuous large-scale events without historical data,the Mean Absolute Percentage Error(MAPE) of the prediction model is within 8%, which shows a high prediction level and can provide decision-making basis for relevant parties.

关 键 词：城市交通 大型活动 PDL模型 客流预测 

分 类 号：U491[交通运输工程—交通运输规划与管理]