Research on the relationship and prediction between Baidu Index and city tourist number—Take Guangzhou City as an example

This paper analyses the relationship between the Baidu index and tourist numbers in Guangzhou, establishing an ARMA model for monthly tourist numbers and making predictions. Seven Baidu keywords, including "Guangzhou food," are added as explanatory variables to build a VAR model for comparison. The results show a long-run equilibrium and Granger causality between tourist numbers and Baidu keywords. The PC-based VAR model's prediction accuracy is 47% higher than the ARMA model. In comparison, the mobile-end VAR model outperforms the PC model by 3% in prediction accuracy but explains 3% more of the variance in tourist number changes. These findings can support the decision-making of relevant departments.