“…Tourist flow data is typical time series data. Existing time series data prediction methods mainly include the following: (a) gray model method (Jiang et al, 2013; Xiao & Duan, 2020); (b) traditional time series modeling methods, such as the autoregressive integrated moving average (ARIMA) model (Calheiros, Masoumi, Ranjan, & Buyya, 2015; Liu, Tian, & Li, 2015; Shukur & Lee, 2015) and random forests (Creamer, 2011; Kusiak, Verma, & Wei, 2013); (c) time series decomposition methods, including spectrum analysis, time series analysis, and Fourier series analysis (Hassani, Soofi, & Zhigljavsky, 2010; Jiang et al, 2013; Tsai, Chen, & Chang, 2016); (d) neural network models, such as the back propagation neural network (BPNN) model (Yang, Li, & Wu, 2017; Zhang, Cui, Feng, Gong, & Hu, 2019) and the recurrent neural network model (Selvin, Vinayakumar, Gopalakrishnan, Menon, & Soman, 2017; Tian & Pan, 2015). The above research methods have great research value for the prediction of tourist flow in scenic spots, but they still have some shortcomings.…”