<![CDATA[This study focuses on Indonesian users' sentiments regarding 9 fintech apps based on their Google Play Store reviews. The rapid growth of the fintech industry in Indonesia makes it crucial to understand user perceptions and satisfaction. Around 2,554 reviews from users of Kredivo, ShopeePay, Dana, GoPay, LinkAja, Bareksa, Flip, Jenius, and OVO were analyzed. The user review text and data were preprocessed using text cleaning, slang normalization, stopword removal, stemming, and the Sastrawi library and were moved through the TF-IDF vectorizer (term frequency-inverse document frequency). The four algorithms were Naive Bayes, Logistic Regression, Support Vector Machine (SVM), and Random Forest. The results showed that SVM (Linear) achieved the best overall balanced performance with an accuracy of 80.23%, precision of 77.79%, recall of 80.23%, and the highest F1-score of 78.53%, outperforming Naive Bayes (accuracy 81.21%, F1-score 78.32%), Logistic Regression (accuracy 80.43%, F1-score 77.81%), and Random Forest (accuracy 78.08%, F1-score 75.81%). While Naive Bayes recorded the highest raw accuracy, SVM was selected as the best model due to its superior F1-score, which provides a more balanced evaluation across all sentiment classes. Machine learning provided a snapshot of the reviews’ sentiments, with 42.4% positive, 51.4% negative, and 6.1% neutral. Kredivo and ShopeePay had the most favorable sentiments of 72.4% and 70.9%. The most salient sentiment indicators include 'bagus' (good) and 'bantu' (help) as top positive classifiers, while 'buruk' (bad) and 'kecewa' (disappointed) emerged as the most prominent negative classifiers, with 'mudah' (easy) and 'cepat' (fast) also strongly associated with positive sentiment. The results of this study give fintech firms a better grasp of user satisfaction, and fintech user positive sentiments.]]>
Copyrights © 2026
