Analisis Sentimen dengan Klasifikasi Naïve Bayes pada Ulasan Tripadvisor di Luxury Resort untuk Strategi Peningkatan Hunian Kamar

Abstrak

The evolution of information technology has transformed how tourists seek and evaluate accommodations, shifting from conventional word-of-mouth to electronic word-of-mouth (eWOM). Leveraging TripAdvisor as a primary platform, this study analyzes guest sentiment toward Amankila Resort using the Naïve Bayes Classifier algorithm. Review data spanning the past five years were obtained through web scraping and processed using Natural Language Processing techniques. Of the four hundred fifty-one reviews, three hundred eighty-eight (eighty-six percent) were categorized as positive, thirty-four (eight percent) as negative, and twenty-nine (six percent) as neutral. The predominance of positive sentiment reflects a robust brand image, with recurring keywords including service, staff, pool, and room. Despite the positive trend, negative feedback provides critical input for service enhancement. This study highlights the importance of encouraging guest reviews to strengthen the resort’s online presence, reputation, and occupancy rate.