基于混频预测模型改进预测精度——以入境旅游为例  

Improving Forecast Accuracy Based on Mixed Frequency Forecasting Model—— A Case Study of Inbound Tourism

作　　者：刘汉[1] 王永莲[1] 

机构地区：[1]吉林大学数量经济研究中心,长春130012

出　　处：《情报杂志》2016年第9期75-79,共5页Journal of Intelligence

基　　金：国家旅游局规划项目(面上项目)"旅游需求监测;预警与中长期趋势的混频实证研究"(编号:15TABG023);教育部人文社会科学研究青年基金项目"宏观经济预测与分析的混频定量研究"(编号:15YJC790055);吉林大学基本科研业务费资助项目"大规模不规则抽样数据的混频因子模型及其应用研究"(编号:2015zz003)研究成果之一

摘　　要：[目的/意义]互联网,尤其是移动互联网已经成为游客获取旅游信息、制定旅游计划、消费旅游产品和反馈旅游体验的主要渠道,网络搜索量在一定程度上反映了这些行为特征。如何充分利用这一丰富的信息提高旅游需求预测精度成为旅游管理的热点问题。[方法/过程]在充分攫取高频搜索引擎数据信息的思想下,以美国来华游客人次为研究对象,开展入境旅游需求的混频预测研究,并与传统同低频预测模型进行比较分析。[结果/结论]实证预测结果表明,基于网络搜索数据的混频预测模型的预测精度相比传统同低频数据模型有了近50%的提升,预测精度和方向上均有显著提高。[ Purpose/Significance ] The Interact, especially mobile Interact, has become the major channel for visitors to obtain the travel information, make travel plans, purchase tourism products and publish their feedbacks. Web search volume can to a certain degree reflect these behavioral characteristics. How to take advantage of the abundant information to improve tourism demand forecasting has become a hot issue in tourism management. ~ Method/Process] In this paper, the mixed frequency model is adopted in inbound tourism demand forecasting with the data of tourists from the Unite States as the sample, under the thought of fully grabbing the high-frequency information of search engine data. [ Result/Conclusion] Empirical results show that the mixed frequency model based on web search data can improve the forecast accuracy by almost 50%, which means significant improvement in forecasting accuracy and direction compared to the traditional low frequency model.

关 键 词：旅游需求 预测精度 混频数据模型 入境游 

分 类 号：F201[经济管理—国民经济] F59