A personalized purchase decision-support model integrating individual expectations and group preferences for hotel selection

Helping consumers make rational purchasing decisions can enhance their satisfaction and loyalty to a hotel platform. This study constructs a personalized hotel decision-making model that integrates individual expectations with group preferences. First, using a questionnaire, we obtain the interval binary expectations of consumers for each attribute and their dynamic risk attitudes. Second, we utilize online reviews to extract group attribute preference weights using Term Frequency-Inverse Document Frequency and word2vec, while obtaining group sentiment expressions via Stanford CoreNLP combined with probabilistic linguistic terminology sets. Next, incorporate the gap between sentiment expression and binary expectations into the value function of prospect theory to obtain individual attribute satisfaction. Finally, we use group preference weights to weight the satisfaction of each attribute, forming the overall satisfaction and generating personalized hotel recommendations. This model is better at capturing consumer group preferences, exerting influence far beyond individual expectations while better meeting personalized needs.