Analisis Sentimen Wisatawan Melalui Data Ulasan Candi Borobudur di Tripadvisor Menggunakan Algoritma Naïve Bayes Classifier

Sentiment analysis of visitors to the tourist destinations Borobudur Temple in Indonesia needs be done determine expected product and service preferences. In addition, sentiment is also helpful for managers adjust tourists infrastructure provided destination area. The classification method used Naïve Bayes Classifier (NBC) against 3850 visitor reviews at Temple. Review data pulled from Tripadvisor web pages filtered by language, review time, travel characteristics analyze foreign traveler preferences comprehensively. This research stage divided into three parts: preparation, processing, analysis, algorithm performance evaluation. SMOTE Upsampling balance data. results implementing obtained an accuracy value 96.36%, a precision 93.23%, recall 100% with Area Under Curve (AUC) 0.714. ranking five famous words show that there are highlights physical condition temple, scenery, visit activities Temple, where four most “temple,” “visit,” “Borobudur,” “sunrise” “place.”