Analisis Sentimen Ulasan Gojek pada Google Play Store Menggunakan Metode Naïve Bayes

Andhika Prasetya (Universitas Bina Sarana Informatika)
Nadif Ramadha (Universitas Bina Sarana Informatika)
Fadhli Firmansyah (Universitas Bina Sarana Informatika)
Jefina Tri Kumalasari (Universitas Bina Sarana Informatika)

Jurnal Nasional Komputasi dan Teknologi Informasi
Vol 8, No 4 (2025): Agustus 2025

Abstrak - Penelitian ini bertujuan untuk menganalisis sentimen pengguna Gojek diGoogle Play Store menggunakan algoritma Naive Bayes Classifier. Analisis sentimen dilakukan untuk memahami persepsi dan tingkat kepuasan pengguna terhadap layanan Gojek berdasarkan ulasan yang mereka bagikan. Data dikumpulkan melalui proses scrapping pada aplikasi Gojek diGoogle Play Store dalam periode tertentu. Tahapan penelitian meliputi preprocessing data (case folding, tokenizing, stopword, dan stemming), pelabelan sentimen (positif, negative, netral), pembobotan kata menggunakan TF-IDF, dan klasifikasi menggunakan algoritma Naive Bayes. Hasil penelitian menunjukkan bahwa dari total 16456 ulasan selanjutnya dilakukan tahapan preprocessing. Maka dihasilkan data sebanyak 11276. Dengan tingkat akurasi klasifikasi mencapai 99.97%. Faktor-faktor yang paling sering dikomentari positif meliputi keandalan driver, keramahan pelayanan, kecepatan proses, dan kualitas layanan yang baik. Sementara keluhan utama terkait dengan masalah teknis dan layanan pelanggan. Penelitian ini memberikan wawasan bagi pengembang Gojek untuk meningkatkan layanan sesuai dengan feedback pengguna.Kata kunci: Analisis Sentimen; Gojek; Naive Bayes Classifier; Abstract - This research aims to analyze Gojek users' sentiment on the Google Play Store using the Naive Bayes Classifier algorithm. Sentiment analysis is conducted to understand users' perceptions and satisfaction levels with Gojek's services based on the reviews they share. Data was collected through the scrapping process on the Gojek application in the Google Play Store over a certain period. The research stages include data preprocessing (case folding, tokenizing, stopword, and stemming), sentiment labeling (positive, negative, neutral), word weighting using TF-IDF, and classification using the Naive Bayes algorithm. The research results show that from a total of 16,456 reviews, the preprocessing stage was subsequently carried out. Thus, a total of 11276 data were produced. With a classification accuracy rate of 99.97%. The factors most frequently commented on positively include driver reliability, service friendliness, process speed, and good service quality. Meanwhile, the main complaints are related to technical issues and customer service. This research provides insights for Gojek developers to improve services based on user feedback.Keywords: Sentiment Analysis; Gojek; Naive Bayes Classifier;