基于数据驱动的旅游需求预测研究  

THE PREDICTION OF TOURISM DEMAND BASED ON DATA DRIVEN

作　　者：胡亚敏 杨力[2] 方润月 HU Ya-min;YANG Li;FANG Run-yue(School of Humanities and Social Science,Huainan,An hui 232001,China;School of Economics and Management,Huainan,An hui 232001,China;School of Mathematics and Big Data,Anhui University of Science&Technology,Huainan,Anhui 232001,China)

机构地区：[1]安徽理工大学人文社会科学学院,安徽淮南232001 [2]安徽理工大学经济与管理学院,安徽淮南232001 [3]安徽理工大学数学与大数据学院,安徽淮南232001

出　　处：《井冈山大学学报（自然科学版）》2022年第4期7-14,共8页Journal of Jinggangshan University (Natural Science)

基　　金：国家社会科学基金重大项目子课题研究项目(20ZDA084)。

摘　　要：合理预测景区客流量不仅可以为景区提供参考,更是旅游治理体系和治理能力现代化建设的内在要求。基于九寨沟风景区官网于2012年5月至2021年5月披露的每日客流量数据,运用Python爬取与九寨沟旅游相关的搜索行为数据和九寨沟每日平均气温,构建ARIMA、SVR模型和加入百度搜索指数与日平均气温的LSTM神经网络模型,对九寨沟风景区客流量进行拟合和预测。结果表明,LSTM神经网络模型预测精度高于ARIMA和SVR模型,加入百度搜索指数和日平均气温的LSTM神经网络模型可以显著提升客流量预测精度。Reasonable prediction of tourist flow in scenic spots can not only provide reference for scenic spots, but also is the inherent requirement of modernization construction of tourism management system and management capacity. Based on the daily passenger flow data disclosed by Jiuzhaigou scenic spot official website from May 2012 to May 2021, the search behavior data related to tourism in Jiuzhaigou and the daily average temperature were extracted, and the passenger flow of Jiuzhaigou scenic spot was fitted and predicted by constructing ARIMA model, SVR model and LSTM neural network model of adding factors. The results show that the prediction accuracy of LSTM neural network model is higher than ARIMA model and SVR model, and the LSTM neural network model with Baidu search index and daily average temperature can significantly improve the prediction accuracy of passenger flow.

关 键 词：旅游需求预测 九寨沟 LSTM神经网络模型 ARIMA模型 SVR模型 

分 类 号：F590.3[经济管理—旅游管理]