Analisis Sentimen Ulasan Aplikasi Shopee Pada Google Play Store Menggunakan Metode FastText

This study aims to analyze user sentiment toward e-commerce services using FastText with the Support Vector Machine (SVM) method. FastText is a machine learning algorithm based on word embedding that can overcome out-of-vocabulary problems through sub-word processing. The dataset contains 5,000 Indonesian-language reviews that have undergone preprocessing and data augmentation to balance the sentiment classes. The model was tested using Linear SVM and Radial Basis Function (RBF) SVM.
The results show that the model with RBF SVM provides the best performance with an accuracy of 82%, precision of 88%, recall of 62%, and F1-score of 60%, while linear SVM achieves an accuracy of 80%, precision of 87%, recall of 59%, and F1-score of 57%.