吉首大学学报(社会科学版)

• 新时代 新青年 新学术:博士生论坛 • 上一篇    下一篇

基于网络搜索指数和EMD-ARIMA-BP组合模型的游客量预测——以张家界为例

陆利军   

  1. (1.中南林业科技大学 旅游学院,湖南 长沙 410004;2.湖南工学院 经济与管理学院,湖南 衡阳 421008;3.衡阳师范学院 湖南省人居环境学研究基地,湖南 衡阳 421008)
  • 出版日期:2019-01-01 发布日期:2019-01-26
  • 作者简介:陆利军,男,中南林业科技大学旅游学院博士研究生,湖南工学院经济与管理学院讲师。
  • 基金资助:
    国家自然科学基金项目(61772192);湖南省人居环境学研究基地开放基金项目(RJ18K03);湖南省高等学校科学研究项目(14C0308)

On the Prediction of Tourist Volume Based on Network Search Index and EMD-ARIMA-BP Combination Model:A Case Study of Zhangjiajie

LU Lijun   

  1. (1.College of Tourism,Central South University of Forestry and Technology,Changsha 410004,China;2.College of Economics and Management,Hengyang Normal University,Hengyang 421008,Hunan China;3.Hunan Human Settlement Environment Research Base,Hengyang Normal University,Hengyang 421008,Hunan China)
  • Online:2019-01-01 Published:2019-01-26

摘要:科学的客流量预测有利于完善旅游安全预警体系和优化旅游资源配置体系。为进一步提高游客量预测的准确度,提出一种基于网络搜索指数的EMD-ARIMA-BP组合模型,以探究互联网时代旅游消费者出行行为规律。该模型首先对网络搜索行为数据进行指数合成,其次利用EMD算法对游客量和网络搜索数据进行去噪处理,最后将ARIMA模型和BP神经网络进行组合,对游客量进行预测。实证分析以张家界为例。研究发现:(1)运用网络搜索数据预测旅游消费者出行行为切实可行,接近于实时的网络数据可以大幅提升预测的时效性;(2)经过EMD去噪算法对游客量与网络搜索行为数据进行去噪处理后,游客量的预测精度有较大程度提高;(3)基于网络搜索指数和EMD-ARIMA-BP组合模型的预测误差显著低于ARIMA模型和BP神经网络等基准模型。

关键词: 网络搜索指数, ARIMA模型, EMD算法, BP神经网络, 游客量预测

Abstract: Scientific prediction of tourist volume is helpful to perfect the early warning system of tourism security and optimize the allocation system of tourism resources.In order to further improve the accuracy of tourist volume prediction,a combination model of EMD-ARIMA-BP neural network based on web search index is proposed to explore the new rules of travel behavior of tourism consumers in the Internet age.The model firstly synthesizes the web search behavior data exponentially,using the EMD algorithm to deal with the noise of the visitor volume and the web search behavior data,combining the econometric prediction model and the BP neural network model to predict tourist volume.The empirical analysis takes the prediction of tourist volume in Zhangjiajie as an example.The results are as follows:(1) it is feasible to predict the travel behavior of tourism consumers by using web search behavior data,and real-time network data can greatly improve the timeliness of prediction;(2) after de-noising the data of tourist volume and web search behavior with EMD de-noising method,the prediction accuracy of tourist volume is improved to a great extent;(3) the prediction error based on the combination of network search index and EMD-ARIMA-BP neural network model is significantly lower than the three benchmark models of ARIMA time series,econometric prediction model and BP neural network.

Key words: network search index, ARIMA model, EMD algorithm, BP neural network;prediction of tourist volume

版权所有 © 2021《吉首大学学报(社会科学版)》编辑部
技术支持:北京玛格泰克科技发展有限公司
公众号 电子书橱 超星期刊 手机浏览 在线QQ