<![CDATA[The rapid development of digital service applications such as Gojek has generated a large number of user reviews on the Google Play Store containing opinions, criticisms, and user experiences. These reviews represent unstructured data that can be utilized to understand user perceptions through sentiment analysis. This study aims to analyze the sentiment of Gojek user reviews using the Long Short-Term Memory (LSTM) method. The research data were collected through web scraping from the Google Play Store, consisting of 1,000 Gojek user reviews. The data underwent several preprocessing stages, including case folding, cleaning, normalization, tokenization, stopword removal, and stemming. Subsequently, the data were transformed into numerical representations, divided into training and testing datasets, and processed using the Long Short-Term Memory (LSTM) method with the assistance of Google Colab. The results showed that, out of 1,000 analyzed reviews, 540 were classified as positive, 270 as negative, and 190 as neutral sentiments. Based on the confusion matrix evaluation, the model achieved an accuracy of 63.00%, a precision of 63.00% for the negative class and 0.00% for the positive class, a recall of 100.00% for the negative class and 0.00% for the positive class, and an F1-score of 77.00% for the negative class and 0.00% for the positive class. These findings indicate that the LSTM method can be applied to classify sentiment in Indonesian-language text. However, the model's performance is still affected by class imbalance, suggesting that further optimization is required to improve classification performance.]]>
Copyrights © 2026
