| Auteur(s) |
Christophe Benavent |
| Type de publication | Working paper |
| Référence | 38th AFM International Congress |
In this paper we pursue two goals: the first is to examine how consumers judge their experience by better understanding how the dimensions of evaluation captured in online reviews affect the rating given using the framework of the bi-factorial theory of satisfaction; the second is methodological, it is to propose an analysis process that combines a topic model to identify evaluation attributes with a GAM regression model to test the effects on satisfaction. The data used represent 24,031 reviews and associated satisfaction ratings left on TripAdvisor concerning 52 hotels in the French Polynesian islands. The results show a dichotomy of the role of experiential and functional attributes on satisfaction and highlight the interest of analyzing customer reviews through NLP methods for hotel managers.
