부산대학교 도서관

In the context of the sharing economy, shared accommodation represented by Airbnb has been widely studied. The study obtains the review data of Airbnb green users in three representative cities of Beijing, Shanghai, and Hong Kong from Inside Airbnb. Ten topics of user reviews are extracted through the Latent Dirichlet Allocation model, and an index system including human, geographic, housing, and environment is constructed. This research uses the sentiment dictionary to calculate the sentiment value of green user reviews, then verifies the interaction effect between each element through multiple regression analysis, and on this basis, uses the Analytic Network Process model to measure the weight of each element. The overall emotional characteristics of green users in Airbnb under the influence of multi-factor interaction are obtained by comprehensive calculation of emotional value and element weight. And the spatial characteristic analysis of the emotional characteristics of green users in Airbnb in three places is carried out. The results show that in Airbnb, green users have a more pronounced emotional tendency towards human and geographical factors during consumption, followed by housing factors, and show lower tendencies towards environmental factors. The research provides a new perspective for optimizing the Airbnb scoring system and promoting the coordinated development of shared accommodation subjects.