Intra-City Tourism Flow Forecasting: A Novel Deep Learning Model

Abstract

Intra-city tourism flow forecasting plays a critical part in urban destination management and planning. However, research on this issue is extremely inadequate because of the challenges of intra-city tourism flow forecasting and the difficulty of obtaining data on intra-city tourism flows. Therefore, this study aims to construct a novel deep learning model that integrates a graph attention network and long short-term memory for the accurate prediction of intra-city tourism flows. A study was conducted in Xiamen, China, to confirm the validity of the proposed model supported by taxi data. The results reveal that the proposed model is applicable to intra-city tourism flow forecasting and outperforms popular benchmarks in terms of forecasting accuracy and robustness. At last, our model effectively obtains information on distribution and temporal fluctuation of tourism flows.