Forecast the recovery trend of inbound tourists based on the ARIMA-LSTM model

This study examines the secondary impact of COVID-19 on the tourism industry, aiming to provide a scientific basis for recovery by predicting inbound tourism demand during the epidemic. Focusing on Beijing, Shanghai, and Guangzhou, data such as the number of inbound tourists and Google search index were used to analyze the spatial and temporal trends before and after the pandemic. The ARIMA-LSTM composite model was employed to forecast inbound tourism for the next two years. Results showed that peak inbound tourism in these cities occurred in April and October, with a 9:1 ratio of traditional to non-traditional tourists. Tourists to Beijing and Shanghai were mostly foreigners, while Guangzhou saw more visitors from Hong Kong and Macao. The ARIMA-LSTM model demonstrated excellent prediction accuracy, with a value exceeding 0.95. Recovery analysis predicts inbound tourism will take one to one and a half years to fully recover across the three cities.