一种基于百度指数的城市日游客规模预测方法  

A prediction method of urban daily tourist scale based on Baidu index

作　　者：任欢 刘婷[1] 康俊锋[3] 潘宁[4] 李敏靓 艾顺毅 REN Huan;LIU Ting;KANG Junfeng;PAN Ning;LI Minliang;AI Shunyi(College of Science,Hangzhou Normal University,Hangzhou 311121,China;College of Resource Environment and Tourism Capital Normal University,Beijing 100048,China;School of Architectural and Surveying&Mapping Engineering,Jiangxi University of Science and Technology,Ganzhou 341000,Jiangxi Province,China;Zhengzhou Tourism College,Zhengzhou 450000,China)

机构地区：[1]杭州师范大学理学院,浙江杭州311121 [2]首都师范大学资源环境与旅游学院,北京100048 [3]江西理工大学建筑与测绘工程学院,江西赣州341000 [4]郑州旅游职业学院,河南郑州450000

出　　处：《浙江大学学报（理学版）》2020年第6期753-761,共9页Journal of Zhejiang University（Science Edition）

基　　金：浙江省大学生科技创新活动计划项目(2018R413033).

摘　　要：百度指数数据为预测游客规模提供了新思路。以杭州市为例,首先研究新浪微博签到数据与统计年鉴中实际游客量的关系,用新浪微博签到人数模拟实际旅游人数,建立杭州市日游客规模自回归移动平均(auto regressive moving average,ARMA)模型,并进行预测;然后结合计量经济学中的协整检验和格兰杰因果关系检验,分析百度指数与微博签到人数之间的关系,利用百度指数空间分布特征及主成分分析后提取的3个解释变量构建向量自回归(vector auto regression,VAR)模型;最后比较了2个模型的预测精度。结果显示,百度指数存在地理空间属性,且与新浪微博签到人数互为格兰杰因果关系,存在1~23 d的滞后期。此外,相比ARMA模型,考虑了百度指数地理属性的VAR模型在样本期内的预测精度提高了13.1%,在样本期外的预测精度提高了27.9%。研究表明,百度指数的时间和空间属性对游客规模预测有重要意义和价值。Baidu index data contains rich information and can be used for tourist scale prediction.Taking Hangzhou as an example,this paper first studies the relationship between Sina Weibo check-in data and the actual number of tourists in the statistical yearbook,and simulates the actual number of tourists with Sina Weibo check-in data,then establishes and forecasts Hangzhou daily tourist scale auto regressive moving average(ARMA)model.Combined with the co-integration theory in econometrics and Granger causality test,the relationship between Baidu index and Weibo check-in data is analyzed.The vector auto regression(VAR)model is constructed by using three explanatory variables extracted based on the spatial distribution characteristics of Baidu index and principal component analysis.Finally,the prediction accuracy of the two models is compared.The results show that Baidu index has geo-spatial attribute,and there is a Granger causality between Baidu index and Sina Weibo check-in data,with a lag of 1 to 23 days.In addition,compared with the ARMA model,the VAR model considering the geographical attributes of Baidu index improves the prediction accuracy by 13.1%during the sample period and 27.9%outside the sample period.This research shows that the temporal and space attributes of Baidu index are of great significance and has great value in the prediction of tourist scale.

关 键 词：百度指数 地理空间属性 ARMA模型 VAR模型 游客规模预测 

分 类 号：F590[经济管理—旅游管理]