Baidu index-based forecast of daily tourist arrivals through rescaled range analysis, support vector regression, and autoregressive integrated moving average.

In the era of big data, Internet search data have become an effective data source for solving social and economic problems, such as tourist arrival prediction. The Internet searches often predate and foretell the actual travels. It is very meaningful to predict the tourist arrivals in a destination based on the Internet search data. Considering the coexistence of linear and nonlinear features of daily tourist arrivals, this paper puts forward a hybrid forecast model of daily tourist arrivals based on Baidu Index, which couples rescaled range analysis (R/S), support vector regression (SVR), and autoregressive integrated moving average (ARIMA). Among them, the SVR excels in nonlinear prediction, and the ARIMA does well in linear prediction. In the hybrid model R/S@SVR-ARIMA, the ARIMA model linearly fits the residual time series predicted by the SVR model, improving the prediction effect. The effectiveness of our model was verified through experiments on Jiuzhai Valley, a famous tourist destination in southwestern China. The research results shed new light on the application of Internet big data in tourism. [ABSTRACT FROM AUTHOR]