Abstract：

 [Purpose/Significance]The Internet, especially mobile Internet, has become the major channel for visitors to obtain the travel information, make travel plans, purchase tourism products, and publish their feedback.Web search volume can to a certain degree reflect these behavioral characteristics. How to take advantage of the abundant information to improve tourism demand forecasting has become a hot issue in tourism management.Method/Process]In this paper, the mixed frequency model is adopted in inbound tourism demand forecasting with the data of tourists from the United States as the sample, under the thought of fully grabbing the high-frequency information of search engine data.[Result/Conclusion]Empirical results show that the mixed frequency model based on web search data can improvethe forecast accuracy by almost 50%, which means significant improvement in forecasting accuracy and direction compared to the tradition-al low-frequency model.

Keywords ：tourism demand forecast accuracy mixed frequency model inbound tourism