<![CDATA[AbstractOnline customer reviews on Online Travel Agent (OTA) platforms serve as digital information sources that shape prospective consumers’ perceptions before hotel booking. This study aims to analyze the sentiment distribution of guest reviews for Le Méridien Bali Jimbaran on Booking.com and examine its potential implications for consumer purchase intention. A quantitative approach was employed through web scraping of 1,002 reviews from March 2023 to February 2026, after data screening and preprocessing, 538 valid reviews were analyzed. Sentiment analysis was conducted using VADER, TF-IDF feature extraction, and Multinomial Naïve Bayes classification. The model achieved an accuracy of 0.75, with an average precision of 0.73, recall of 0.59, and F1-score of 0.60. The results show that positive sentiment dominates at 67.84%, followed by neutral sentiment at 22.12% and negative sentiment at 10.04%. These findings indicate that guest perceptions are generally positive, although several service aspects still require improvement. Consumer purchase intention was not directly measured but discussed conceptually based on guest perceptions reflected in OTA reviews. This study provides managerial insights for strengthening digital reputation and supporting the potential formation of consumer purchase intention.Keywords: Sentiment Analysis; Online Travel Agent; Purchase Intention; Naïve Bayes; Booking.comAbstrakUlasan tamu pada platform Online Travel Agent (OTA) menjadi sumber informasi digital yang membentuk persepsi calon konsumen sebelum melakukan pemesanan hotel. Penelitian ini bertujuan untuk menganalisis distribusi sentimen ulasan tamu Le Méridien Bali Jimbaran pada Booking.com serta mengkaji potensi implikasinya terhadap minat beli konsumen. Penelitian ini menggunakan pendekatan kuantitatif melalui teknik web scraping terhadap 1.002 ulasan periode Maret 2023-Februari 2026, setelah seleksi dan pra-pemrosesan, diperoleh 538 ulasan valid untuk dianalisis. Analisis sentimen dilakukan menggunakan VADER, ekstraksi fitur TF-IDF, dan klasifikasi Multinomial Naïve Bayes. Model memperoleh akurasi 0,75, precision 0,73, recall 0,59, dan F1-score 0,60. Hasil menunjukkan dominasi sentimen positif sebesar 67,84%, diikuti sentimen netral 22,12% dan negatif 10,04%. Temuan ini menunjukkan bahwa persepsi tamu cenderung positif, meskipun masih terdapat aspek layanan yang perlu diperbaiki. Minat beli tidak diukur langsung, tetapi dibahas secara konseptual berdasarkan persepsi dalam ulasan OTA. Hasil penelitian memberikan masukan manajerial untuk memperkuat reputasi digital dan mendukung potensi pembentukan minat beli konsumen.Kata Kunci: Analisis Sentimen; Online Travel Agent; Minat Beli; Naïve Bayes; Booking.com]]>
