基于百度指数时空分布的旅游趋势预测研究——以上海市为例  

Tourism Trend Prediction Based on Baidu Index Spatial and Temporal Distribution

作　　者：康俊锋[1] 郭星宇 方雷 KANG Jun-Feng;GUO Xing-Yu;FANG Lei(School of Architecture and Surveying Engineering,Jiangxi University of Science and Technology,Ganzhou Jiangxi 341000,China;Department of Environmental Science and Engineering,Fudan University,Shanghai 200433,China)

机构地区：[1]江西理工大学建筑与测绘工程学院,江西赣州341000 [2]复旦大学环境科学与工程系,上海200433

出　　处：《西南师范大学学报（自然科学版）》2020年第10期72-81,共10页Journal of Southwest China Normal University(Natural Science Edition)

基　　金：国家重点研发计划项目(2016YFC0803105);国家留学基金资助项目(201808360065);江西省教育厅科学技术研究项目(GJJ150661);国家自然科学基金青年基金资助项目(41701462)。

摘　　要：科学、准确、便捷、低成本地预测旅游趋势对提高景区的科学管理能力及避免因旅游人数过多导致的公共安全问题具有重要意义.研究选取2011-2018年中国各省级行政区(港澳台除外)与上海市旅游相关的百度指数数据和上海市国内游客数据构建旅游趋势预测模型.通过Granger因果检验、 ARIMA模型挖掘公众网络搜索行为与现实旅游行为的映射关系;依据百度指数数据的时空分布规律,采用支持向量机方法对百度指数数据进行聚类,解决不同省份百度指数因变化趋势近似而造成的多重共线问题,优化后的预测模型平均预测精度提升23.36%.研究发现:(1)昨天的搜索者就是今天的旅游者;(2)基于地理位置的旅游空间距离与旅游出游率呈反比、百度指数的地理位置属性有助于提升预测精度.In this paper,the Baidu index and the number of Chinese domestic tourists(1.38 billion in total)of each consecutive monthly travel destination of Shanghai from 2011 to 2018 have been studied.Through the Granger causality test,ARIMA model,spatial clustering method and principal component analysis,the mapping relationship between Internet virtual space and the real world has been explored.With the help of Spatio-temporal distribution pattern analysis and seasonal trend analysis,the multicollinearity problem of a similar time trend of different sources has been solved,thus the average prediction accuracy of the optimized prediction model been increased by 23.35%.Moreover,it is concluded that“yesterday s searchers are today s tourists”,“travel distance is inversely proportional to travel rate”and“the geographical location attribute of the search index is helpful to improve the prediction accuracy”.Tourism forecast can provide scientific and accurate decision-making basis for the scenic spot management department to ensure the safety of the scenic spot and tourism experience.

关 键 词：百度指数 旅游预测 时空分布 ARIMA模型 支持向量聚类 地理信息系统 

分 类 号：F59[经济管理—旅游管理]