Perbandingan Analisis Sentimen Ulasan Pengguna Aplikasi Shopeepay pada Google Play Store Menggunakan Metode Support Vector Machine (SVM) dan Decision Tree Classification

Abstract

Technological developments, especially in finance, have led to the rise of financial technology (fintech), including digital wallets like Shopeepay, which facilitate electronic transactions via smartphones. This study compares Support Vector Machine (SVM) and Decision Tree classification in classifying sentiments of Shopeepay user reviews from the Google Play Store using Support Vector Machine (SVM) achieved 95% accuracy, with 95% precision, 94% recall, and 94% F1-score for negative reviews, and 95% precision, 96% recall, and 96% F1-score for positive reviews. Decision Tree Classification obtained 91% accuracy, with lower recall for negative reviews 82% due to overfitting on the majority class and sensitivity to data imbalance. While Decision Tree Classification offers high interpretability, Support Vector Machine (SVM) is more reliable and consistent for sentiment classification of ShopeePay user reviews.