Sentiment Classification of Public Tweets Towards CGV Cinemas on Social Media X Using Naive Bayes Algorithm

Abstract

In the era of digital communication, sentiment analysis on social media platforms provides businesses with valuable insights into public perception. This research aims to classify public sentiment toward CGV cinemas in Indonesia through tweets collected from Social Media X using the Naive Bayes algorithm. A total of 4,000 tweets were preprocessed through a series of text normalization techniques, including tokenization, stop word removal, and stemming. Text features were transformed using the TF-IDF method. The Naive Bayes classifier was trained and evaluated using an 80:20 train-test split. Experimental results showed an overall classification accuracy of 38.05%, with the model performing significantly better on positive sentiments (F1-score: 0.538) than on neutral and negative ones. These findings highlight the capability and limitations of traditional probabilistic classifiers when dealing with short, noisy textual data in multilingual social contexts. This study contributes to applied sentiment analysis and offers a baseline for future comparison with more sophisticated models