Topic Modeling LDA and SVM in Sentiment Analysis of Hotel Reviews

Abstract

The number of visitor comment review data that enters the TripAdvisor and Agoda sites continues to grow over time, this makes it difficult for the hotel to obtain overall information from all comment reviews. Therefore, the purpose of this study is to apply topic modeling and classifying in the analysis of hotel service sentiment.  The data for comment reviews were obtained from 3 five-star hotels, namely 1-HTL, 2-HTL and 3-HTL. The hotel has a five-star rating and has the most comments compared to other hotels in Jakarta. The topic modeling method using Latent Dirichlet Allocation (LDA) in this study succeeded in dividing the comments into several topics that were often discussed from Indonesian and English comments regarding the hotel services provided. By using Support Vector Machine (SVM) obtained the number of positive, negative and neutral comments.