ARIMA-LSTM model based on least square weighting to predict number of inbound tourists: A case study of Shanghai

doi:10.3785/j.issn.1008-9497.2023.04.014

Journal of Zhejiang University (Science Edition)

2023, Vol. 50

Issue (4): 508-520 DOI: 10.3785/j.issn.1008-9497.2023.04.014

Tourism

ARIMA-LSTM model based on least square weighting to predict number of inbound tourists: A case study of Shanghai

Junfeng KANG¹(

),Yue FU¹,Lei FANG²(

),Mimi LI³,Yujing XIE²,Chaoyang ZHOU⁴

^1.School of Civil and Surveying & Mapping Engineering，Jiangxi University of Science and Technology，Ganzhou 341000，Jiangxi Province，China
^2.Department of Environmental Science and Engineering，Fudan University，Shanghai 200433，China
^3.School of Hotel and Tourism Management，The Hong Kong Polytechnic University，Hong Kong 999077，China
^4.Jiangxi Provincial Defense Science and Technology Information and Satellite Application Center，Nanchang 330036，China

Download:

HTML( 3 )

PDF(4684KB)
Export: BibTeX | EndNote (RIS)

Abstract

In order to reduce the secondary impact of the COVID-19 on the tourism industry, accurate prediction of the demand of inbound tourism market during the epidemic period can provide a scientific basis for later recovery and development of tourism. Taking Shanghai as the study area, the number of inbound tourists, major source countries, Google search index, confirmed cases of the epidemic and other data were selected to quantitatively analyze the spatial characteristics and the temporal trend of inbound tourism before and after the epidemic, the ARIMA-LSTM combination model weighted by the least square method was used to predict the number of inbound tourists after the epidemic. The results show that: (1) before and after the outbreak of the epidemic, the Asian tourists occupied the core position of the inbound tourism market, and the proportion of traditional inbound tourists and non-traditional inbound tourists is about 9∶1; (2) the number of inbound tourists demonstrates a long-term positive correlation and Granger causality with the Google search index, but there is no significant correlation with the confirmed cases of the epidemic; (3) by comparing the model evaluation indicators, it is found that when the $R 2$ value of ARIMA-LSTM model is higher than 0.8, the model fits well, and the prediction error is smaller than that of a single model, and the prediction accuracy is higher, which means that the model can be uniformly applied to the recovery prediction of tourist numbers before, during and after the epidemic; (4) the number of inbound tourists from 2021-2024 is predicted, and it shows that the tourism trend during this period presents an obvious U-shaped change. After the comprehensive release of the epidemic in December 2022, the tourism industry began to gradually recover, and it is expected that the number of inbound tourists will return to the preepidemic tourism level by the end of 2024, that is, the recovery period is about one and a half years.

Key words： COVID-19 Shanghai tourism forecast ARIMA-LSTM model least squares method Google index

Received: 22 August 2022 Published: 17 July 2023

CLC:

F 590

Corresponding Authors: Lei FANG E-mail: junfeng.kang@jxust.edu.cn;fanglei@fudan.edu.cn

	Service
	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Junfeng KANG
	Yue FU
	Lei FANG
	Mimi LI
	Yujing XIE
	Chaoyang ZHOU

Cite this article:

Junfeng KANG,Yue FU,Lei FANG,Mimi LI,Yujing XIE,Chaoyang ZHOU. ARIMA-LSTM model based on least square weighting to predict number of inbound tourists: A case study of Shanghai. Journal of Zhejiang University (Science Edition), 2023, 50(4): 508-520.

URL:

https://www.zjujournals.com/sci/EN/Y2023/V50/I4/508

基于最小二乘法赋权的ARIMA-LSTM模型预测入境旅游人数——以上海市为例

为降低新冠病毒感染疫情大流行对旅游业的二次冲击，对疫情防控期间入境旅游市场的需求进行准确预测可为后期旅游业复苏提供科学依据。以上海市为研究区域，选取入境旅游人数、主要客源国、谷歌搜索指数、新增确诊病例数等数据，定量分析疫情前后入境旅游人数的空间变化特征及时间变化趋势，并用基于最小二乘法赋权的ARIMA-LSTM模型预测疫情后的入境旅游人数。结果表明：（1）疫情发生前后，亚洲客源市场一直占据入境旅游市场的核心地位，且传统入境游客与非传统入境游客的比例约为9∶1；（2）入境旅游人数与谷歌搜索指数存在长期正相关及格兰杰因果关系，与确诊病例数无明显相关性；（3）通过对比模型评价指标发现，当ARIMA-LSTM模型的 $R 2$ 大于0.8时，拟合较好，预测误差较单一模型小，预测精度较单一模型高，适用于疫情前、中、后期的旅游人数恢复预测；（4）对2021—2024年入境旅游人数进行恢复预测，发现该期间入境旅游人数呈明显的U形曲线。自2022年12月疫情全面放开后，旅游业逐步恢复，预计入境旅游人数在2024年12月恢复至疫情前水平，即需1.5 a的恢复期。

关键词： 新冠病毒感染, 上海旅游预测, ARIMA-LSTM模型, 最小二乘法, 谷歌搜索指数

Table 1 Statistics on the number of inbound tourists in Shanghai and the Google index

Table 2 Epidemic-related data statistics

Fig.1 LSTM structure

Fig.2 ARIMA-LSTM model frame diagram

Fig.3 Changes of inbound tourists from different source countries in Shanghai

Fig.4 Proportion of major source countries of inbound tourists in Shanghai from 2017 to 2020

Fig.5 Composition of the number of tourists entering Shanghai from four continents

Fig.6 Google search keyword filter

Fig.7 Comparison of the trend of the number of inbound tourists in Shanghai and the Google search index （normalized） from 2017 to 2021

Fig.8 Comparison of the trend of the number of confirmed cases and imported cases abroad

Fig.9 Correlation coefficient between the number of inbound tourists in Shanghai and the keyword Google search index and newly confirmed cases

Table 3 Granger causality test results

Fig.10 Effect of different models on predicting the number of inbound tourists before and after the epidemic

Table 4 Evaluation results of prediction accuracy of different models before and after the epidemic

Fig.11 Forecast results of the number of inbound tourists in Shanghai from 2022 to 2024


[1]	任欢，刘婷，康俊锋. 一种基于百度指数的城市日游客规模预测方法［J］. 浙江大学学报（理学版）， 2020， 47（6）： 753-761. DOI：10.3785/j.issn.1008-9497. 2020.06.014 REN H， LIU T， KANG J F. A method for predicting the scale of daily tourists in cities based on Baidu index［J］. Journal of Zhejiang University （Science Edition）， 2020， 47（6）： 753-761. DOI：10.3785/j.issn.1008-9497.2020.06.014 doi: 10.3785/j.issn.1008-9497.2020.06.014

[2]	卢璐，孙根年. 2008年至2018年我国大陆地区入境旅游的危机周期及市场归因［J］. 浙江大学学报（理学版）， 2021， 48（3）： 377-390. DOI：10.3785/j.issn. 1008-9497.2021.03.014 LU L， SUN G N. The crisis cycle and market attribution of inbound tourism in mainland country from 2008 to 2018［J］. Journal of Zhejiang University（Science Edition）， 2021， 48（3）： 377-390. DOI：10.3785/j.issn.1008-9497.2021.03.014 doi: 10.3785/j.issn.1008-9497.2021.03.014

[3]	SHARMA G D， THOMAS A， PAUL J. Reviving tourism industry post-COVID-19： A resilience-based framework［J］. Tourism Management Perspectives， 2021， 37： 100786. DOI：10.1016/j.tmp.2020.100786 doi: 10.1016/j.tmp.2020.100786

[4]	VOLGGER M， TAPLIN R， AEBLI A. Recovery of domestic tourism during the COVID-19 pandemic： An experimental comparison of interventions［J］. Journal of Hospitality and Tourism Management， 2021， 48： 428-440. DOI：10.1016/j.jhtm.2021.07.015 doi: 10.1016/j.jhtm.2021.07.015

[5]	ZHANG H， SONG H， WEN L， et al. Forecasting tourism recovery amid COVID-19［J］. Annals of Tourism Research， 2021， 87（4）： 103149. DOI：10. 1016/j.annals.2021.103149 doi: 10. 1016/j.annals.2021.103149

[6]	KUMAR A， MISRA S C， CHAN F T S. Leveraging AI for advanced analytics to forecast altered tourism industry parameters： A COVID-19 motivated study［J］. Expert Systems with Applications， 2022， 210： 118628. DOI：10.1016/j.eswa.2022.118628 doi: 10.1016/j.eswa.2022.118628

[7]	SONG H， QIU R T R， PARK J. A review of research on tourism demand forecasting： Launching the annals of tourism research curated collection on tourism demand forecasting［J］. Annals of Tourism Research， 2019， 75： 338-362. DOI：10.1016/j.annals.2018.12.001 doi: 10.1016/j.annals.2018.12.001

[8]	GHU F L. A fractionally integrated autoregressive moving average approach to forecasting tourism demand［J］. Tourism Management， 2008， 29（1）： 79-88. DOI：10.1016/j.tourman.2007.04.003 doi: 10.1016/j.tourman.2007.04.003

[9]	JANGHEE C， SEUNG C D， LEE T H. Forecasting tourism demand of Jeju Island using GAM and ARMA［J］. Korean Management Consulting Review， 2018， 18（2）： 187-194.

[10]	CHEN J， HUANG M， FU J. Comparison of China PR inbound tourism forecast methods-ARIMA-based model， BP neural network model and BP-ARIMA mixed model［J］. Basic & Clinical Pharmacology & Toxicology， 2020， 127： 96-96.

[11]	DEININGER M， KOELLNER T， BREY T， et al. Towards mapping and assessing Antarctic marine ecosystem services：The Weddell sea case study［J］. Ecosystem Services， 2016， 22： 174-192. DOI：10. 1016/j.ecoser.2016.11.001 doi: 10. 1016/j.ecoser.2016.11.001

[12]	AYDIN M. The impacts of political stability， renewable energy consumption， and economic growth on tourism in Turkey： New evidence from Fourier Bootstrap ARDL approach［J］. Renewable Energy， 2022， 190： 467-473. DOI：10.1016/j.renene.2022. 03.144 doi: 10.1016/j.renene.2022. 03.144

[13]	CHATZIANTONIOU I， FILIS G， EECKELS B， et al. Oil prices， tourism income and economic growth： A structural VAR approach for European Mediterranean countries［J］. Tourism Management， 2013， 36： 331-341. DOI：10.1016/j.tourman.2012. 10.012 doi: 10.1016/j.tourman.2012. 10.012

[14]	LAW R， LI G， FONG D K C， et al. Tourism demand forecasting： A deep learning approach［J］. Annals of Tourism Research， 2019， 75： 410-423. DOI：10.1016/j.annals.2019.01.014 doi: 10.1016/j.annals.2019.01.014

[15]	ZHANG Y， TANG Z. PSO-weighted random forest for attractive tourism spots recommendation［J］. Future Generation Computer Systems， 2022， 127： 421-425. DOI：10.1016/j.future.2021.09.029 doi: 10.1016/j.future.2021.09.029

[16]	HONG W C， DONG Y， CHEN L Y， et al. SVR with hybrid chaotic genetic algorithms for tourism demand forecasting［J］. Applied Soft Computing， 2011， 11（2）： 1881-1890. DOI：10.1016/j.asoc. 2010.06.003 doi: 10.1016/j.asoc. 2010.06.003

[17]	FAN G F， JIN X R， HONG W C. Application of COEMD-S-SVR model in tourism demand forecasting and economic behavior analysis： The case of Sanya city［J］. Journal of the Operational Research Society， 2022， 73（7）： 1474-1486. DOI：10.1080/01605682.2021.1915192 doi: 10.1080/01605682.2021.1915192

[18]	TEIXEIRA J P， FERNANDES P O. Tourism time series forecast different ANN architectures with time index input［J］. Procedia Technology， 2012， 5： 445-454. DOI：10.1016/j.protcy.2012.09.049 doi: 10.1016/j.protcy.2012.09.049

[19]	TASYUREK M， CELIK M. RNN-GWR： A geographically weighted regression approach for frequently updated data［J］. Neurocomputing， 2020， 399： 258-270. DOI：10.1016/j.neucom.2020.02.058 doi: 10.1016/j.neucom.2020.02.058

[20]	FENG L， HAO Y K. Optimization algorithm of tourism security early warning information system based on long short-term memory （LSTM）［J］. Computational Intelligence and Neuroscience， 2021， 2021： 9984003. DOI：10.1155/2021/9984003 doi: 10.1155/2021/9984003

[21]	MO K C， SHIN S H， HLEE S， et al. Online tourism review：Three phases for successful destination relationships［J］. Asia Pacific Journal of Information Systems， 2015， 25（4）： 746-762. DOI：10.14329/apjis.2015.25.4.746 doi: 10.14329/apjis.2015.25.4.746

[22]	SHERAFATIAN-JAHROMI R， OTHMAN M S， LAW S H， et al. Tourism and CO2 emissions nexus in Southeast Asia： New evidence from panel estimation［J］. Environment Development and Sustainability， 2017， 19（4）： 1407-1423. DOI：10.1007/s10668-016-9811-x doi: 10.1007/s10668-016-9811-x

[23]	ASLANARGUN A， MAMMADOV M， YAZICI B， et al. Comparison of ARIMA， neural networks and hybrid models in time series： Tourist arrival forecasting［J］. Journal of Statistical Computation and Simulation， 2007， 77（1）： 29-53. DOI：10.1080/10629360600564874 doi: 10.1080/10629360600564874

[24]	CHEN K Y. Combining linear and nonlinear model in forecasting tourism demand［J］. Expert Systems with Applications， 2011， 38（8）： 10368-10376. DOI：10. 1016/j.eswa.2011.02.049 doi: 10. 1016/j.eswa.2011.02.049

[25]	LEE J. A reformulation of weighted least squares estimators［J］. American Statistician， 2009， 63（1）： 49-55. DOI：10.1198/tast.2009.0011 doi: 10.1198/tast.2009.0011

[26]	SUN C， JI S. The least squares estimator of random variables under sublinear expectations［J］. Journal of Mathematical Analysis and Applications， 2017， 451（2）： 906-923. DOI：10.1016/j.jmaa.2017.02.020 doi: 10.1016/j.jmaa.2017.02.020

[27]	LI X， LAW R， XIE G， et al. Review of tourism forecasting research with internet data［J］. Tourism Management， 2021： 83： 104245. DOI：10.1016/j.tourman.2020.104245 doi: 10.1016/j.tourman.2020.104245

[28]	SUN S， WEI Y， TSUI K L， et al. Forecasting tourist arrivals with machine learning and internet search index［J］. Tourism Management， 2019， 70： 1-10. DOI：10.1016/j.tourman.2018.07.010 doi: 10.1016/j.tourman.2018.07.010

[29]	YANG Y， FAN Y， JIANG L， et al. Search query and tourism forecasting during the pandemic： When and where can digital footprints be helpful as predictors？［J］. Annals of Tourism Research， 2022， 93： 103365. DOI：10.1016/j.annals.2022.103365 doi: 10.1016/j.annals.2022.103365

[30]	WANG Y， GUO Y. Forecasting method of stock market volatility in time series data based on mixed model of ARIMA and XGBoost［J］. China Communications， 2020， 17（3）： 205-221. DOI：10. 23919/jcc.2020.03.017 doi: 10. 23919/jcc.2020.03.017

[31]	YU Y， SI X， HU C， et al. A review of recurrent neural networks： LSTM cells and network architectures［J］. Neural Computation， 2019， 31（7）： 1235-1270. DOI：10.1162/neco_a_01199 doi: 10.1162/neco_a_01199

[32]	FRAME J M， KRATZERT F， RANEY A， et al. Post-processing the national water model with long short-term memory networks for streamflow predictions and model diagnostics［J］. Journal of the American Water Resources Association， 2021， 57（6）： 885-905. DOI：10.1111/1752-1688.12964 doi: 10.1111/1752-1688.12964

[33]	唐弘久，保继刚. 我国主要入境客源地游客的时空特征及影响因素［J］. 经济地理， 2018， 38（9）： 222-230. DOI：10.15957/j.cnki.jjdl.2018.09.026 TANG H J， BAO J G. The spatiotemporal characteristics and influencing factors of tourists from the main inbound tourist sources in my country［J］. Economic Geography， 2018， 38（9）： 222-230. DOI：10.15957/j.cnki.jjdl.2018.09.026 doi: 10.15957/j.cnki.jjdl.2018.09.026

[34]	张国平，刘晓鹰. 基于旅游目的分组的城镇居民国内旅游消费构成演变趋势探讨［J］. 商业时代， 2014 （2）： 31-33. DOI：10.3969/j.issn.1002-5863.2014.02.012 ZHANG G P， LIU X Y. Discussion on the evolution trend of domestic tourism consumption composition of urban residents based on tourism purpose grouping［J］. The Age of Business， 2014（2）： 31-33. DOI：10. 3969/j.issn.1002-5863.2014.02.012 doi: 10. 3969/j.issn.1002-5863.2014.02.012

[35]	JIN X C， QU M， BAO J. Impact of crisis events on Chinese outbound tourist flow： A framework for post-events growth［J］. Tourism Management， 2019， 74： 334-344. DOI：10.1016/j.tourman.2019.04.011 doi: 10.1016/j.tourman.2019.04.011

[36]	AREF F. Sense of community and participation for tourism development［J］. Life Science Journal-Acta Zhengzhou University Overseas Edition， 2011， 8（1）： 20-25.

[37]	KIM H J， CHEN M H， JANG S S. Tourism expansion and economic development： The case of Taiwan［J］. Tourism Management， 2006， 27（5）： 925-933. DOI：10.1016/j.tourman.2005.05.011 doi: 10.1016/j.tourman.2005.05.011

[1]	Qinling YAN,Peiyu LIU. Impact of meteorological factors and control measures on the spread of COVID-19 epidemic[J]. Journal of Zhejiang University (Science Edition), 2023, 50(2): 144-152.

[2]	ZHAO Lining, LI Junyi. A study on the influence of residential environment on anxiety of urban residents during COVID-19 epidemic[J]. Journal of Zhejiang University (Science Edition), 2021, 48(5): 642-650.

[3]	XU Tingting, LI Gang, GAO Xing, WANG Jiaobei, WANG Yu, ZHANG Qianxi. Spatio-temporal evolution and influencing factors of COVID-19 epidemic in Zhejiang province[J]. Journal of Zhejiang University (Science Edition), 2021, 48(3): 356-367.

[4]	ZHANG Xiaodong, HAN Haoying, CHEN Yu, LI Xiancan, LUO Guona. A Study on the characteristics and influencing factors of hospital spatial distribution in China[J]. Journal of Zhejiang University (Science Edition), 2021, 48(1): 84-92.

[5]	Ye XingdeJiang Jinsheng. A New Pseudospectral Approximationfor. the Biharmonic Boundary Value Problem[J]. Journal of Zhejiang University (Science Edition), 1995, 22(1): 13-21.

Viewed

Full text

Abstract

Cited

Shared

Discussed