Sentiment Analysis of Indonesian Tourism Social Media Using Naïve Bayes for Decision Support

Abstract

The tourism sector remains one of the main contributors to Indonesia's economy, but its development is still challenged by inadequate communication and publication. Understanding visitors' opinions is essential for improving tourist destinations. This study aims to analyze public sentiment toward Indonesian tourist attractions by automatically processing visitor responses using sentiment analysis. The proposed approach applies text mining with the Naïve Bayes algorithm to classify sentiments efficiently. Data were collected through the X platform API using tourism-related keywords and hashtags, providing real-time public opinions on Indonesian destinations. A web-based application was developed using Python to visualize the sentiment analysis results as graphs showing the distribution of sentiment categories. The proposed model achieved an accuracy of 87%, demonstrating its effectiveness in classifying public responses. The findings provide useful insights for tourism stakeholders to evaluate visitor perceptions, identify areas for improvement, and support data-driven decision-making to enhance tourism services and visitor experiences. This study contributes to tourism research by applying the Naïve Bayes algorithm to Indonesian-language tweets related to tourist attractions, demonstrating the feasibility of text mining for automatically analyzing public opinion and supporting tourism development.