Perbandingan Kinerja Support Vector Machine Dan Random Forest Untuk Klasifikasi Sentimen Pengguna Aplikasi Gojek Dengan Optimasi Smote

Rihastuti, Siti
Rosyidi, Afnan

Algoritme Jurnal Mahasiswa Teknik Informatika
Vol 5 No 3 (2025): Oktober 2025 || Algoritme Jurnal Mahasiswa Teknik Informatika

This study compares the performance of Support Vector Machine (SVM) and Random Forest in classifying Gojek user sentiment using 2,000 Indonesian-language reviews (1,351 positive, 566 negative, 83 neutral). After data preprocessing and TF-IDF feature extraction, SMOTE was applied to balance the training data in each fold. Using Stratified K-Fold Cross-Validation, results showed that Random Forest achieved higher and more consistent accuracy (84.1%) than SVM (76.1%). The Paired t-test and McNemar’s Test (p-value < 0.05) confirmed that the Random Forest’s superiority was statistically significant. Overall, both models were effective, but Random Forest performed better for Gojek sentiment classification, supporting user satisfaction monitoring and complaint detection.