Sentiment analysis and topic modelling on user-generated content in hospitality: a case study on customer perceptions in Ho Chi Minh City, Vietnam

This study leverages sentiment analysis and topic modelling to enhance understanding of customer perceptions in the hospitality industry, through an empirical analysis of user-generated content from Booking.com related to establishments in Ho Chi Minh City, Vietnam. By employing latent Dirichlet allocation, the research uncovers critical themes impacting customer satisfaction, including service quality, the overall hotel experience, location convenience, dining options, and room comfort and cleanliness. The effectiveness of various machine learning (ML) and deep learning (DL) models is evaluated for sentiment analysis. The convolutional neural network model, in particular, demonstrates superior performance with an accuracy of 0.95 and an F1-score of 0.97. Highlighting the application of sophisticated ML and DL techniques to analyse complex patterns in user-generated feedback, this research study offers valuable insights into brand equity and customer experiences within the hospitality sector.