A novel BEMD-based method for forecasting tourist volume with search engine data

Abstract

As helpful big data, search engine data (SED) regarding tourism-related factors have currently been introduced to tourist volume prediction, but they have been shown to impact the tourism market on different timescales (or frequency band). This study develops a novel forecasting method using an emerging multiscale analysis—bivariate empirical mode decomposition (BEMD)—to investigate multiscale relationships. Three major steps are performed: (1) SED process to construct an informative index from sufficient SED using statistical analyses, (2) multiscale analysis to extract scale-aligned common factors from the bivariate data of tourist volumes and SED using BEMD, and (3) tourist volume prediction using an SED-based index. In the empirical study, the novel BEMD-based method with SED is used to forecast the tourist volume of Hainan in China, a global tourist attraction, and significantly outperforms both popular techniques (not considering SED or multiscales) and similar variants (considering SED or multiscales) in accuracy and robustness.

Seasonal and trend forecasting of tourist arrivals: An adaptive multiscale ensemble learning approach

“…The decomposition technology based on frequency domain analysis has been proven to show superior performance in processing nonlinear and high‐frequency time series, Shao et al (2017) comprehensively reviewed the decomposition methods used in electricity demand forecasting. Although the application fields are different, in tourism demand forecasting, the decomposition methods widely applied as apart from the combined forecasting model including at least five methods: (1) Fourier transfer (Apergis et al, 2017; Hu, 2021); (2) Wavelet decomposition (Kummong & Supratid, 2016); (3) Empirical mode decomposition (Chen et al, 2012; Li & Law, 2020; Tang et al, 2020); (4) Singular spectrum analysis (Beneki et al, 2012; Hassani et al, 2015) and (5) Filtering analysis.…”

“…Hu (2021) applied a Fourier series to remove noise in tourism demand data, proposed a gray forecasting model that offered high forecasting performance. Tang et al (2020) introduced the bivariate empirical mode decomposition to investigate the relationship between the tourism‐related factors and tourist volume forecasting of Hainan in China. Xie et al (2020) applied the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) into tourism arrivals forecasts, the empirical results demonstrated its superiority.…”

Decomposition Methods for Tourism Demand Forecasting: A Comparative Study

“…In recent years, the AI-based forecasting models, as a popular and effective tool for processing the nonlinear data without any assumptions such as stationarity or distribution, have already been applied to tourism forecasting (Song, Qiu, and Park 2019;Sun et al 2019). Generally, the common AI models include support vector regression (SVR), the fuzzy time series, genetic algorithms, back propagation neural network (BPNN), extreme learning machine (ELM), random vector functional link network (RVFL), and long-/short-term memory (LSTM) (S. Li et al 2018;Sun et al 2019;Song, Qiu, and Park 2019;Tang et al 2020). For example, Sun et al (2019) proposed a forecasting approach with kernel ELM for significantly improving the forecasting performance.…”

“…RVFL. RVFL is an extended version of single-hidden layer feed-forward neural networks, without tuning the weights and biases iteratively (Tang et al 2020). The RVFL network structure connects the input layer to the hidden layer, and also directly connects the input layer to the output layer, which can be represented as…”

“…Generally, the common AI models include support vector regression (SVR), the fuzzy time series, genetic algorithms, back propagation neural network (BPNN), extreme learning machine (ELM), random vector functional link network (RVFL), and long-/short-term memory (LSTM) (S. Li et al 2018; Sun et al 2019; Song, Qiu, and Park 2019; Tang et al 2020). For example, Sun et al (2019) proposed a forecasting approach with kernel ELM for significantly improving the forecasting performance.…”