The Application of Multinomial Naive Bayes for Sentiment Analysis of Cultural Tourism Reviews: A Case Study of Borobudur and Prambanan Temples

A comparative sentiment analysis of tourist reviews from Borobudur and Prambanan Temples, two UNESCO World Heritage Sites, reveals significant differences in visitor perceptions. Prambanan recorded a higher proportion of positive sentiment (82.5%) compared to Borobudur (73.7%), with recurring negative complaints centered on heat, ticket prices, and parking. The study employed a hybrid methodology combining web scraping, automated labeling via VADER, preprocessing, TF-IDF feature weighting, and Multinomial Naive Bayes classification, achieving accuracies of 81.9% for Prambanan and 74.7% for Borobudur. Visualization techniques like word clouds further highlighted key pain points, offering actionable insights for tourism management to enhance service quality based on data-driven feedback.

Alfaritzi, I., Abirawa, B., Javid Ahmedthian, S., Wisnubroto, M. S., Farid, F., & Untoro, M. C. (2026). THE APPLICATION OF MULTINOMIAL NAIVE BAYES FOR SENTIMENT ANALYSIS OF CULTURAL TOURISM REVIEWS: A CASE STUDY OF BOROBUDUR AND PRAMBANAN TEMPLES. Antivirus : Jurnal Ilmiah Teknik Informatika, 20(1), 40–47. https://doi.org/10.35457/dj70zr92