Mitigating the disturbances of events on tourism demand forecasting

Developing an accurate tourism forecasting decision support system can help the tourism department achieve optimal resource allocation, which is crucial for achieving sustainable tourism operation management and a circular economy. Recent decades have witnessed the frequent strikes of crisis events and mega-events, which profoundly influence tourist arrival volume and bring a great challenge to forecasting tourist arrival volume. To solve this issue, we develop a deep learning framework to forecast the tourist arrival volume utilizing search engine data containing the trends of tourism intention and different event information. Our proposed model is novel for the following reasons: (1) The disturbance value can predict tourist arrival volume in coordination with the trend of travel plans. (2) Compared with the traditional models, our model can reduce the complexity of the model while maintaining accuracy. (3) Our proposed framework introducing event-related search volumes can capture the concerns of tourists and the potential loss of tourist arrivals, enhancing the model’s predictive power. Experimental results show that our model can accurately forecast the tourist arrival volume by employing the monthly data in Beijing and Sanya, China. Moreover, our findings provide policymakers with more understanding of the relationship between various predictive factors and tourist arrivals. Based on the forecasting results, allocating an appropriate amount of clean energy transportation capacity, garbage treatment capacity, and fresh food supply capacity to the city can effectively promote the circular economy.