Tourist arrival forecasting using feed search information

The feed index is a weighted sum of the number of reactions (i.e. reading, comments, retweets, likes and dislikes, and so on.) that the content engine actively recommends and distributes to the users. It provides valuable information from big data on the Internet and high marketing value to Destination Marketing Organization as the content can be customized. A large-scale empirical study on the impact of the feed index on tourist arrival forecasting accuracy has been conducted, with a new approach proposed to incorporate the feed index into the tourist arrival forecasting model with higher forecasting accuracy. Firstly, the empirical results suggest that the feed index for different keywords reflects varying tourist preferences and has different impacts on tourist arrival movements, with variant lead-lag relationships. Secondly, the study shows that keywords need to be carefully selected based on theoretical analysis plus new methods such as entropy analysis. Therefore, it is proposed that entropy is employed to select the keywords and time lags, thus helping improve forecasting accuracy. © 2023 Informa UK Limited, trading as Taylor & Francis Group

The feed index is a weighted sum of the number of reactions (i.e. reading, comments, retweets, likes and dislikes, and so on.) that the content engine actively recommends and distributes to the users. It provides valuable information from big data on the Internet and high marketing value to Destination Marketing Organization as the content can be customized. A large-scale empirical study on the impact of the feed index on tourist arrival forecasting accuracy has been conducted, with a new approach proposed to incorporate the feed index into the tourist arrival forecasting model with higher forecasting accuracy. Firstly, the empirical results suggest that the feed index for different keywords reflects varying tourist preferences and has different impacts on tourist arrival movements, with variant lead-lag relationships. Secondly, the study shows that keywords need to be carefully selected based on theoretical analysis plus new methods such as entropy analysis. Therefore, it is proposed that entropy is employed to select the keywords and time lags, thus helping improve forecasting accuracy.