Big Data-based Sentiment Analysis on TripAdvisor Reviews Using Naïve Bayes Classification

This research aims to analyze the sentiment of tourist reviews on the TripAdvisor platform towards a luxury resort in Bali by utilizing the Naïve Bayes classification method. The review data is analyzed to identify positive, negative, and neutral sentiments. Three variants of Naïve Bayes algorithm (GaussianNB, MultinomialNB, and BernoulliNB) were implemented and evaluated for performance. The results showed that the GaussianNB model provided the highest classification accuracy of 0.89. Further analysis revealed that the model effectively identified positive sentiments, but had challenges in classifying negative and neutral sentiments. Word cloud visualization confirmed the focus of positive reviews on aspects of accommodation, service and facilities, which can serve as a reference for the hospitality industry. This study concludes that big data-based sentiment analysis is an important tool for understanding customer perceptions, noting the need for further model development to improve the identification of minority sentiments.