Although user-generated content (UGC) in the digital economy has garnered substantial attention for economic forecasting, the impact of marketer-generated content (MGC) on tourism demand has been empirically established but less explored in combination with UGC. Utilizing daily visitor data from the Palace Museum in China as a case study, this research examines this understudied relationship and its role in forecasting visitor numbers. We formulate and evaluate six specific machine-learning models using big data collected from two prominent social media platforms: TikTok and Weibo. To enhance model interpretability, we introduce a novel methodology grounded in Shapley regression and weighted contribution coefficients, providing a quantitative assessment of the individual impacts of MGC and UGC. Empirical results reveal that MGC serves as a leading predictor for daily visitor arrivals at the Palace Museum, and optimal accuracy is achieved when both MGC and UGC are integrated into the predictive algorithm. This study fills an existing gap in the literature and offers both theoretical and practical contributions to tourism demand forecasting.
Introduction
Interpretable Forecasting Framework and Methods
Empirical Research
Literature review
Methods
Data Collection
Data Analysis
Findings
Gratification at pre-travel stage
Gratification at Post-Travel Stage
Gratification at Post-Travel Stage
Discussion and Conclusion
Limitations and Future Research Directions
Reference
(0)
(0)