Forecast of the Tourist Volume of Sanya City by XGBoost Model and GM Model

Sanya is a famous tourist city in China. Considering the monthly tourist data form Sanya Tourism, Culture, Radio, Television and Sports Bureau from 2012 to 2018, this paper adopts XGB and GM learning models to analyze tourist volume of Sanya City in the past seven years. Forecast accuracy of these two models are therefore compared. The fitting results indicate that the mean absolute percentage error (MAPE) of XGB model is 3.92% while the GM model achieves MAPE of 3.18%. Both prediction models reach a high accuracy. Based on the fitting result of these two models, a combined model is further proposed to forecast the tourist volume, the number of overnight transit tourists, and the tourist income in Sanya City for 2019 and 2020, respectively. The forecast results indicate that: (i) the tourist volume will reach 35,560,100 and 37,938,200 in 2019 and 2020 respectively; (ii) the number of overnight transit tourists will reach 875,200 and 1,027,000 in 2019 and 2020 respectively; and (iii) the tourist income will reach RMB 55.335 billion and RMB 64.44 billion in 2019 and 2020 respectively. Based on the forecasted tourist volume, recommendations are proposed to improve the city's tourist reception capacity and service capacity such that high quality of satisfaction can be provided for tourists in an accordance with the level of an international tourist city.