Analyzing tourism reviews using an LDA topic-based sentiment analysis approach
Twil Ali, Omar Bencharef, Soulaimane Kaloun
Abstract
It has become increasingly necessary to automate systems for organizing and classifying user reviews according to their domain-specific aspects and sentiment polarities, as online customer opinions have increased on specialized platforms and social networks. This study's methodology employs a combination of topic modeling and sentiment analysis, as well as human validation techniques of topic labels, to extract valuable insights about Marrakech city from TripAdvisor reviews. Through this technique, tourism practitioners and field specialists may dive deeper into online users generated content, leveraging aspect-based sentiment analysis to explore each destination's weaknesses and strengths.•Data collection and pre-processing.•Extracting latent topics using LDA algorithm (Latent Dirichlet Allocation) on collected reviews.•Applying sentiment analysis to each topic.