摘要
The prediction of the Baidu index for tourism demand has been increasingly focused on by scholars. However, few studies have evaluated the predictive power of the Baidu index for hotel guest arrivals in fine granularity at the micro level. Taking Guilin as a case study, we use the OLS regression method to quantitatively investigate the forecasting power of the Baidu index for daily hotel guest arrivals and to comprehensively evaluate the performance of the forecasting model and to optimize the forecasting model by deeply mining the hidden characteristics of tourism flow in a special case study. The contributions of this paper mainly have threefold: first, to the best of our knowledge, based on the actual full-example of daily hotel guest check-in data in fine granularity, we evaluated the predictive power of the Baidu index by comparison of 5 forecasting models for the first time. Second, we proposed two metrics for forecasting: the trend forecasting index and the forecasting stability index. Finally, we introduce a kind of punishment strategy to optimize forecasting models based on the potential pattern of research objects.
学科与关键词
- Forecasting Tourist Arrivals with Machine Learning and Internet Search IndexSun, Shaolong; Wang, Shouyang; Wei, Yunjie*; Yang, Xianduan; Tsui, Kwok-LeungIEEE International Conference on Big Data (IEEE Big Data), Boston, MA, 2017-12-11 To 2017-12-14.被引 3
- Tourist arrival forecasting using feed search informationHe, Kaijian; Yang, Qian; Wu, Don; Zou, Yingchao*Current Issues in Tourism, 2023.
- Integration of Internet search data to predict tourism trends using spatial-temporal XGBoost composite modelKang, Junfeng; Guo, Xingyu; Fang, Lei*; Wang, Xiangrong; Fan, ZhengqiuInternational Journal of Geographical Information Science, 2022, 36(2): 236-252.被引 10
