Causal evaluation of intangible cultural heritage effects on tourism development with double machine learning

This study utilizes panel data from 30 provinces and municipalities in China to construct a data system for intangible cultural heritage. The research employs dual machine learning models and random forest models to visually present the influencing variables of China’s intangible cultural heritage tourism. Thereafter, mediating variables were introduced to analyze the nonlinear predictive feature importance on intangible cultural heritage tourism, visualized through forest plots and dependence plots. The results show that (1) most of the 13 factors that affect intangible cultural heritage have a positive relationship with it. However, growth slows or plateaus after reaching a certain development threshold. (2) The cultural stations exert the strongest positive effect on intangible cultural heritage tourism. (3) Infrastructure and the service sector—including star-rated hotels, industry personnel, and travel agencies—have a positive, albeit lagging, impact on its development.