Perbandingan TF-IDF dan Sentence-IndoBERT untuk Klasifikasi Emosi Ulasan Produk Indonesia
DOI:
https://doi.org/10.35889/jutisi.v15i4.3872Keywords:
, emotion classification, Indonesian product reviews, text representation, logistic regression, machine learningAbstract
Informal Indonesian product reviews containing lexical variation posed challenges for emotion classification. This study aimed to compare term frequency-inverse document frequency and Sentence-IndoBERT representations using the same logistic regression classifier so that the effect of text representation could be examined under controlled conditions. The PRDECT-ID dataset was audited, cleaned of duplicates and label conflicts, and reduced to 5,261 reviews. The data were stratified into 70% training, 15% validation, and 15% testing sets. Performance was evaluated using accuracy, macro precision, macro recall, macro F1-score, weighted F1-score, and a confusion matrix. The results showed that term frequency-inverse document frequency achieved 64.05% accuracy and a 60.88% macro F1-score, outperforming Sentence-IndoBERT, which achieved 63.67% accuracy and a 59.28% macro F1-score. The novelty lay in a controlled experiment using the same classifier, data partition, and evaluation procedure. The findings indicated that word-based representation was more effective and computationally efficient for the evaluated dataset.
Keywords: Emotion Classification; Indonesian Product Reviews; Text Representation; Logistic Regression; Machine Learning
Abstrak
Ulasan produk berbahasa Indonesia yang singkat, informal, dan mengandung variasi kata telah menimbulkan kesulitan dalam klasifikasi emosi. Penelitian ini telah membandingkan representasi term frequency-inverse document frequency dan Sentence-IndoBERT menggunakan regresi logistik yang sama agar pengaruh representasi teks dapat diamati secara terkontrol. Dataset PRDECT-ID telah diaudit, dibersihkan dari duplikasi dan konflik label, lalu menghasilkan 5.261 ulasan yang dibagi secara terstratifikasi menjadi 70% data pelatihan, 15% data validasi, dan 15% data pengujian. Kinerja telah dievaluasi menggunakan akurasi, presisi makro, recall makro, skor F1 makro, skor F1 berbobot, dan confusion matrix. Hasil pengujian telah menunjukkan bahwa term frequency-inverse document frequency memperoleh akurasi 64,05% dan skor F1 makro 60,88%, lebih tinggi daripada Sentence-IndoBERT dengan akurasi 63,67% dan skor F1 makro 59,28%. Kebaruan penelitian telah ditunjukkan melalui eksperimen terkontrol dengan pengklasifikasi, pembagian data, dan prosedur evaluasi yang sama. Temuan tersebut telah memperlihatkan bahwa representasi berbasis kata lebih efektif dan efisien untuk dataset yang digunakan.
Kata kunci: Klasifikasi Emosi; Ulasan Produk Indonesia; Representasi Teks; Logistic Regression; Machine Learning
References
[1] R. Sutoyo, S. Achmad, A. Chowanda, E. W. Andangsari, dan S. M. Isa, “PRDECT-ID: Indonesian product reviews dataset for emotions classification tasks,” Data Brief, vol. 44, pp. 108554, Okt 2022, doi: 10.1016/J.DIB.2022.108554.
[2] A. D. P. Ariyanto, Fari Katul Fikriah, dan Arif Fitra Setyawan, “Emotion Detection Using Contextual Embeddings for Indonesian Product Review Texts on E-commerce Platform,” Pixel :Jurnal Ilmiah Komputer Grafis, vol. 17, no. 1, pp. 179–185, Jul 2024, doi: 10.51903/pixel.v17i1.2010.
[3] V. B. Lestari dan C. A. Hutagalung, “Evaluation of TF-IDF Extraction Techniques in Sentiment Analysis of Indonesian-Language Marketplaces Using SVM, Logistic Regression, and Naive Bayes,” J-KOMA Journal of Computer Science and Applications, vol. 8, no. 1, pp. 22–2025, 2025, doi: 10.21009/j.
[4] R. D. Kurniawan, “GoPay App Review Sentiment Classification Optimization Using a Combination of Text Representation and Machine Learning,” 2024. [Daring]. Tersedia pada: http://ejournal.uksw.edu/ijiteb
[5] S. Aras, M. Yusuf, R. Ruimassa, E. A. B. Wambrauw, dan E. B. Palalangan, “Sentiment Analysis on Shopee Product Reviews Using IndoBERT,” ISI Journal of Information System and Informatics, vol. 6, no. 6, pp. 132-141, 2024, Diakses: 25 Juni 2026. [Daring]. Tersedia pada: https://doi.org/10.51519/journalisi.v6i3.814
[6] K. C. Pradhisa dan R. Fajriyah, “Analisis Sentimen Ulasan Pengguna E-commerce di Google Play Store Menggunakan Metode IndoBERT,” Building of Informatics, Technology and Science (BITS), vol. 6, no. 1, pp. 68-79, Jun 2024, doi: 10.47065/bits.v6i1.5247.
[7] K. Wau, “Application of Fine-Tuned IndoBERT for Sentiment Classification Local Product Reviews on Tokopedia Marketplace with Limited Dataset,” Journal of Artificial Intelligence and Engineering Applications (JAIEA), vol. 5, no. 1, pp. 2808–4519, 2025, [Daring]. Tersedia pada: https://ioinformatic.org/
[8] A. S. Rizky dan E. Y. Hidayat, “Emotion Classification in Indonesian Text Using IndoBERT,” Computer Engineering and Applications, vol. 14, no. 1, pp. 2252–4274, 2025.
[9] J. I. T. Krisna, A. Luthfiarta, L. D. Cahya, S. Winarno, dan A. Nugraha, “Comparing Optimizer Strategies For Enhancing Emotion Classification In IndoBERT Models,” Advance Sustainable Science, Engineering and Technology, vol. 6, no. 2, pp. 12-21, Feb 2024, doi: 10.26877/asset.v6i2.18228.
[10] F. Basbeth dan D. H. Fudholi, “Klasifikasi Emosi Pada Data Text Bahasa Indonesia Menggunakan Algoritma BERT, RoBERTa, dan Distil-BERT,” JURNAL MEDIA INFORMATIKA BUDIDARMA, vol. 8, no. 2, pp. 1160, Apr 2024, doi: 10.30865/mib.v8i2.7472.
[11] Y. A. Singgalen, “Performance Analysis of IndoBERT for Sentiment Classification in Indonesian Hotel Review Data,” Journal of Information System Research (JOSH), vol. 6, no. 2, pp. 976–986, Jan 2025, doi: 10.47065/josh.v6i2.6505.
[12] A. Devi, P. Ariyanto, F. Katul Fikriah, dan A. F. Setyawan, “Impact of Statistical and Semantic Features Extraction for Emotion Detection on Indonesian Short Text Sentences,” CommIT (Communication and Information Technology) Journal, vol. 19, no. 1, pp. 1-13. 2025.
[13] M. Habib Algifari dan D. Nugroho, “Emotion Classification of Indonesian Tweets using BERT Embedding,” Journal of Applied Informatics and Computing, vol. 7, no. 2, pp. 172-176. 2023. [Daring]. Tersedia pada: http://jurnal.polibatam.ac.id/index.php/JAI
[14] H. Hartatik dan A. Syafrianto, “Penerapan Model Sentence-BERT Untuk Sistem Rekomendasi Buku Berbasis Konten Di Perpustakaan Digital,” Jurnal Dialektika Informatika (Detika), vol. 6, no. 1, pp. 12–19, Nov 2025, doi: 10.24176/detika.v6i1.15916.
[15] S. Pradeepa dkk., “HGATT_LR: transforming review text classification with hypergraphs attention layer and logistic regression,” Sci. Rep., vol. 14, no. 1, pp. 69-77, Des 2024, doi: 10.1038/s41598-024-70565-6.
[16] K. Kowsari, K. J. Meimandi, M. Heidarysafa, S. Mendu, L. Barnes, dan D. Brown, “Text classification algorithms: A survey,” 2019, MDPI AG. doi: 10.3390/info10040150.
[17] M. Khadhraoui, H. Bellaaj, M. Ben Ammar, H. Hamam, dan M. Jmaiel, “Survey of BERT-Base Models for Scientific Text Classification: COVID-19 Case Study,” Applied Sciences (Switzerland), vol. 12, no. 6, pp. 23-31, Mar 2022, doi: 10.3390/app12062891.
[18] A. Gasparetto, M. Marcuzzo, A. Zangari, dan A. Albarelli, “ Survey on Text Classification Algorithms: From Text to Predictions,” Information (Switzerland), vol. 13, no. 2, pp. 112-123, Feb 2022, doi: 10.3390/info13020083.
[19] A. Vabalas, E. Gowen, E. Poliakoff, dan A. J. Casson, “Machine learning algorithm validation with a limited sample size,” PLoS One, vol. 14, no. 11, pp. 281-292, Nov 2019, doi: 10.1371/journal.pone.0224365.
