<![CDATA[Sentiment analysis is used to explore user perceptions of fintech services such as Tunaiku through the evaluation of customer reviews. This study specifically aims to compare the performance of several sentiment classification algorithms to determine the most optimal model for classifying Tunaiku app user reviews. The dataset used in this study is a collection of Tunaiku app user reviews obtained from the Google Play Store, with a total of 18,458 reviews. This study compares the performance of five classification algorithms, namely Naïve Bayes, Logistic Regression, Support Vector Machine (SVM), XGBoost, and a hybrid SVM-XGBoost model. The research stages include text preprocessing, feature extraction using TF-IDF, and the application of a validated classification model using the cross-validation method. Model performance evaluation is carried out based on accuracy, precision, recall, and F1-score metrics. The test results showed that Naïve Bayes (91.96%), Logistic Regression (92.81%), SVM (92.56%), and XGBoost (92.52%) provided good performance, while the hybrid SVM-XGBoost model produced the best performance with the highest accuracy of 93.05%. These findings indicate that the hybrid approach is more effective in analyzing user review sentiment and has the potential to be a basis for decision-making in improving Tunaiku's service quality according to user needs.]]>
Copyrights © 2025
