基于混频预测模型改进预测精度——以入境旅游为例

发布时间：2018-06-09 12:57

本文选题：旅游需求 + 预测精度　；参考：《情报杂志》2016年09期

【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。
[Abstract]:[purpose / meaning] the Internet, especially the mobile Internet, has become a major channel for tourists to access tourism information, plan tourism, consume tourism products and feedback tourism experiences, The amount of network search reflects these behavioral characteristics to some extent. How to make full use of this abundant information to improve the accuracy of tourism demand prediction has become a hot issue in tourism management. [method / process] based on the idea of fully grabbing the high-frequency search engine data information, the mixed frequency prediction of inbound tourism demand was carried out with the American tourists coming to China as the research object, and compared with the traditional low-frequency prediction model. [results / conclusion] empirical prediction results show that the prediction accuracy of the hybrid prediction model based on the network search data is nearly 50% higher than that of the traditional low-frequency data model, and the prediction accuracy and direction are significantly improved.
【作者单位】：吉林大学数量经济研究中心;
【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
【分类号】：F224;F592

【参考文献】