Sentiment Analysis of the Film "JUMBO" on Twitter Using the Naive Bayes Method and Support Vector Machine (SVM) with a Text Mining Approach

Abstract

This study aims to perform sentiment analysis on reviews of the film “JUMBO” collected from the Twitter platform, using the Naive Bayes and Support Vector Machine (SVM) methods. The data were gathered through a crawling process on Twitter, yielding 2,011 tweets, which were then processed through several pre-processing steps, including case folding, cleaning, normalization, tokenization, stopword removal, and stemming. Subsequently, the data were transformed into numerical representations using TF-IDF, followed by sentiment labeling into positive, negative, and neutral categories. For the Naive Bayes method, training and evaluation were conducted using 5-fold Cross Validation. The results showed that the Naive Bayes model achieved an accuracy of 80.60%, precision of 73.83%, recall of 73.50%, and an F1-score of 69.98%. Meanwhile, the SVM method obtained an accuracy of 75.87%, precision of 76.36%, recall of 62.45%, and an F1-score of 65.64%. Compared to the baseline random classifier, which only achieved an accuracy of 32.47%, both primary methods significantly outperformed it in classifying film review sentiments. The analysis also indicates that the F1-score is lower than the accuracy due to the imbalanced data distribution, with a considerably higher number of positive reviews. This study also presents visualizations of sentiment distribution and word clouds to provide a clearer understanding of audience opinions. The results demonstrate that the Naive Bayes method performs well and has potential for use in sentiment analysis of films on social media platforms. These findings are expected to provide valuable insights for the creative industry, particularly in evaluating audience responses and improving the quality of future film productions.