Sentiment Analysis on Indonesian Telegram Reviews Using Naïve Bayes, SVM, Random Forest, and Boosting Models

This paper presents a comprehensive sentiment analysis of user reviews of the Telegram application in Indonesia, employing various machine learning models such as Naïve Bayes, Support Vector Machine (SVM), Random Forest, XGBoost, and LightGBM. Considering the growing reliance on user-generated reviews to inform decision-making in digital platforms, this research addresses the urgent need for accurate and scalable sentiment analysis methods. 50,000 Google Play Store reviews were collected and preprocessed. We performed feature extraction using TF-IDF and n-grams to convert text into numerical formats suitable for analysis. The SVM model achieved the highest accuracy at 93.4%. The analysis identified significant positive and negative sentiments expressed by users, providing valuable insights into the Indonesian language.