Forecasting Tourism Demand by a Novel Multi-Factors Fusion Approach

Abstract

The volatility of tourism demand is often caused by some irregular events in recent years. Typically, inbound tourists are quite sensitive to various factors, including the exchange rate fluctuation, consumer price index, personal or household income or consumption expenditure. We combine these multivariate time series data onto an ingenious multi-factor fusion strategy to contribute to precise tourism demand forecasting. A novel hybrid deep learning forecasting approach is developed by integrating several modules such as improved complete ensemble empirical mode decomposition with adaptive noise, intrinsic mode functions classification, multi-factors fusion and predictors matching. The monthly tourist flow data of Shanghai inbounding from USA, Korea and Japan are conducted to verify the performance of the proposed approach, which outperforms all benchmark models for different prediction horizons. The experimental results show that introducing external influencing factors can improve the prediction accuracy significantly, and therefore confirm the rationality and validity of the proposed approach.

A Systematic Review of the Literature on the Use of Artificial Intelligence in Forecasting the Demand for Products and Services in Various Sectors

“…Another important model within this group is the one proposed by the document [28] which converts time series into images and then uses special convolutional networks for processing. Finally, two papers from the group [29] and [30] focus their models on the decomposition of the original time series into several component series to then find the best forecasting methods for each of them. Notable in this sense is the document [30] that proposes using statistical or simple machine learning methods, such as ARIMA or SVR, for low complexity components and using neural networks with bidirectional GRU for the forecast of high complexity components.…”

“…Nine documents also raise the difficulties caused by the fact that the factors that affect demand are very diverse and numerous, such as calendar factors, climatic factors, economic factors, market factors, etc. Which makes the construction of adequate models extremely challenging [12], [13], [17], [29], [30], [26], [34], [37], [41].…”

“…The demand for many products exhibits temporal patterns, such as trend and seasonality, however, other patterns are superimposed on these temporal patterns due to external factors such as the economy, climate, competition, etc. Simultaneously modeling both types of patterns can be a very complex task and a great challenge for forecasting models [16], [20], [29], [28], [30], [26], [40], [41].…”

“…Another problem that significantly affects forecasts is disruptive events, such as calendar events or COVID 19. The robustness of www.ijacsa.thesai.org forecast models with respect to these types of events is highly desirable and is addressed by documents [13], [30], [37] and [41].…”

“…Simple Machine Learning Methods Within this group are the traditional machine learning methods multiple linear regression (MLR) and support vector machine (SVM). Twelve documents propose the use of these methods, however they are proposed in combination with other more complex methods [12], [15], [18]- [21], [29], [30], [35], [40] or with classical statistical methods [17].…”