摘要
文章基于干预ARIMA模型对2014年1月至2022年2月海南省游客接待人数序列进行了建模与预测,考虑疫情对旅游业的显著冲击,分别对疫情前后阶段的序列特征进行分析:针对疫情前具有季节性波动的序列,构建ARIMA(0,1,1)×(0,1,0)_(12)模型,并将其外推至疫情期间,以估计无干预情形下的游客接待水平。通过实际值与预测值的比较,提取疫情干预效应,并通过白噪声检验验证干预效应显著存在。随后,采用脉冲响应函数与阶跃函数刻画疫情的短期与长期影响,构建最终预测模型,该模型的MAPE为10.6%,预测性能良好,尤其能捕捉到2020年2月游客人数的骤降和随后整体回升的趋势。
The paper models and forecasts the sequence of tourist receptions in Hainan Province from January 2014 to February 2022 based on the intervention ARIMA model,and analyzes the characteristics of the sequence in the pre-epidemic situation and post-epidemic situation phases respectively,considering the significant impact of the epidemic situation on the tourism industry.For the pre-epidemic situation sequence with seasonal fluctuations,the ARIMA(0,1,1)×(0,1,0)_(12)model is constructed,and it is extrapolated to the epidemic situation period to estimate the level of tourist receptions in the no-intervention scenario.The epidemic situation intervention effect is extracted by comparing the actual values with the forecast values,and the white noise test verifies the significant presence of the intervention effect.Subsequently,the impulse response function and the step function are used to characterize the short-term and long-term impacts of the epidemic situation,and the final forecast model is constructed.The model has a MAPE of 10.6%and good forecast performance,especially capturing the sudden drop in tourist numbers in February 2020 and the subsequent overall recovery trend.
作者
曾柏桐
ZENG Botong(School of Mathematical Sciences,South China Normal University,Guangzhou 510631,China)
出处
《现代信息科技》
2025年第14期78-83,共6页
Modern Information Technology
