<![CDATA[The rapid advancement of digital technology has increased the use of public service applications such as PLN Mobile. However, numerous user reviews highlight technical issues that may reduce customer satisfaction. This study aims to compare the performance of Naïve Bayes and Long Short-Term Memory (LSTM) algorithms in classifying user review sentiments on the PLN Mobile application. The dataset was collected from Google Play Store using Google Play Scraper. The research involved text preprocessing (case folding, tokenizing, stopword removal, stemming, TF-IDF), oversampling with SMOTE, and evaluation using accuracy, precision, recall, and F1-score metrics. The results show that Naïve Bayes achieved 96% accuracy, while LSTM reached 98%. LSTM also outperformed Naïve Bayes in all evaluation metrics, indicating superior capability in capturing contextual and sequential patterns in textual data. Therefore, LSTM is more recommended for sentiment analysis of complex text data.]]>
Copyrights © 2026
