<![CDATA[This study examines sentiment analysis of user reviews of the Stockbit application obtained from the Google Play Store and platform X (formerly Twitter). The aim of this research is to classify user opinions into two sentiment categories: positive and negative, using the Support Vector Machine (SVM) method. A total of 3,000 review data points were used in this study, consisting of 2,100 training data points and 900 test data points (stratified split with a 70:30 ratio) to ensure balanced sentiment distribution. The research process includes text preprocessing, feature weighting using Term Frequency-Inverse Document Frequency (TF-IDF), sentiment classification with the SVM algorithm, and model performance evaluation. Based on the evaluation results, the SVM model demonstrated high performance with an accuracy of 95.5%, precision of 93.5%, recall of 97.4%, and an F1-score of 95.3%. Although its accuracy is lower than that of Maulana et al.'s (2024) study, which achieved 99.50% on the Pluang application, this research excels in using data from two different platforms and evaluating class imbalance, making the analysis results more representative of real-world conditions. These findings indicate that SVM remains an effective method for text-based sentiment analysis in digital financial service applications.]]>
Copyrights © 2025
