Article Sidebar
Main Article Content
Abstract
User reviews of the Disney+ Hotstar application on the Google Play Store present a variety of sentiments, particularly concerning the paid subscription feature. This study aims to analyze these sentiments using the Naïve Bayes classification method, categorizing user opinions into positive, negative, and neutral classes. A total of 30,571 Indonesian- language reviews were collected through web scraping, followed by a preprocessing phase that included case folding, stopword removal, and stemming. The Term Frequency-Inverse Document Frequency (TF- IDF) technique was applied to weight the significance of words. The dataset was split into 80% training and 20% testing portions. The classification model achieved an accuracy of 78%, showing reliable performance in identifying sentiment patterns. However, performance on the neutral class was lower, indicating room for improvement through better preprocessing or class balancing. The findings provide insights for Disney+ Hotstar to better understand user perceptions and guide enhancements to the subscription service.