Analisis Performa Metode BiLSTM untuk POS Tagging pada Kalimat Bahasa Indonesia Tingkat Dasar
-
Dhafa Widy Pratama
*
Universitas Teknologi Yogyakarta, Daerah Istimewa Yogyakarta,
Indonesia
- Ledy Elsera Astrianty Universitas Teknologi Yogyakarta, Daerah Istimewa Yogyakarta, Indonesia (*) Corresponding Author
Abstract
The effect of hyperparameters on part-of-speech tagging (POS Tagging) performance for the Indonesian language in deep learning models is rarely examined in isolation, because earlier studies usually add complex components (pretrained embeddings, CNN, CRF, or BERT), so the effect of hyperparameters on a plain Bidirectional Long Short-Term Memory (BiLSTM) cannot be isolated. The novelty of this study is a stepwise analysis of five hyperparameters (embedding dimension, hidden dimension, learning rate, train-test split ratio, and number of epochs) on a three-layer BiLSTM built from scratch in NumPy, without pretrained embeddings or CRF, on basic-level Indonesian sentences, that is, sentences commensurate with the comprehension level of elementary school (SD) students, represented by first-grade SD/MI textbooks. Secondary data comprised 10,747 sentences with 23 labels, combining the IDN Tagged Corpus and 747 textbook sentences. The effect of hyperparameters proved highly uneven: learning rate and hidden-layer capacity determined performance, whereas embedding dimension, split ratio, and number of epochs shifted accuracy by less than 0.5 percentage points. The best configuration reached 95.92% accuracy and 0.9966 weighted ROC-AUC, with labels X, UH, and FW being the hardest to predict. A five-seed test of the final configuration showed stable results (accuracy 95.82 ± 0.08% at 50 epochs and 95.76 ± 0.17% at 98 epochs; the difference between the two was not significant, p = 0.26), but small differences between hyperparameter settings have not been tested statistically. This study provides a simple and transparent baseline and a priority order for hyperparameter tuning as a starting point for NLP practitioners.
Downloads
References
Af’idah, D. I., Dairoh, D., Handayani, S. F., Pratiwi, R. W., & Sari, S. I. (2022). Sentimen Ulasan Destinasi Wisata Pulau Bali Menggunakan Bidirectional Long Short Term Memory. MATRIK : Jurnal Manajemen, Teknik Informatika dan Rekayasa Komputer, 21(3), 607–618. https://doi.org/10.30812/matrik.v21i3.1402
Akbar, M. I., Heriansyah, R., & Irfani, M. H. (2026). Pengaruh Splitting Data terhadap Akurasi Klasifikasi Demam Berdarah Dengue Menggunakan K-Nearest Neighbors. Jurnal Algoritme, 6(2), 244–256. https://doi.org/10.35957/algoritme.v6i2.13274
Alfian, M., Yuhana, U. L., & Siahaan, D. (2023). Indonesian Part-of-Speech Tagger: A Comparative Study. 2023 10th International Conference on Advanced Informatics: Concept, Theory and Application, ICAICTA 2023. https://doi.org/10.1109/ICAICTA59291.2023.10390353
Alfian, M., Yuhana, U. L., Siahaan, D., Munazharoh, H., & Pardede, E. (2025). LFF-POS: A linguistic fusion method to handle out-of-vocabulary words in low-resource part-of-speech tagging. MethodsX, 15. https://doi.org/10.1016/j.mex.2025.103615
Amalia, J., Pakpahan, J., Pakpahan, M., & Panjaitan, Y. (2022). Model Klasifikasi Berita Palsu Menggunakan Bidirectional LSTM dan Word2vec sebagai Vektorisasi. JATISI (Jurnal Teknik Informatika dan Sistem Informasi), 9(4), 3319–3331. https://doi.org/10.35957/jatisi.v9i4.1332
Aritonang, Y. V., Napitupulu, D. P., Sinaga, M. H., & Amalia, J. (2022). Pengaruh Hyperparameter pada Fasttext terhadap Performa Model Deteksi Sarkasme Berbasis Bi-LSTM. JATISI (Jurnal Teknik Informatika dan Sistem Informasi), 9(3), 2612–2625. https://doi.org/10.35957/jatisi.v9i3.1331
Asnawiyah, & Eka Putra, R. (2025). Perbandingan Algoritma LSTM dan BiLSTM Untuk Analisis Sentimen Multi-Class Media Sosial Twitter. Journal of Informatics and Computer Science (JINACS), 6(03), 778–786. https://doi.org/10.26740/jinacs.v6n03.p778-786
Astari, Y., Afiyati, & Rozaqi, S. W. (2021). Analisis Sentimen Multi-Class pada Sosial Media menggunakan metode Long Short-Term Memory (LSTM). Jurnal Linguistik Komputasional, 4(1), 8–12. https://doi.org/10.26418/jlk.v4i1.43
Bui, N. T., Savova, G., & Wang, L. (2025). Assessing the Macro and Micro Effects of Random Seeds on Fine-Tuning Large Language Models. http://arxiv.org/abs/2503.07329
Cahyo, P. W., Aesyi, U. S., Setianto, W. A., & Sulaiman, T. (2025). A Novel Named Entity Recognition approach of Indonesian fake news using part of speech and BERT model on presidential election. International Journal of Information Management Data Insights, 5(2), 100354. https://doi.org/10.1016/j.jjimei.2025.100354
Dewayani, S. (2023). Bahasa Indonesia Aku Bisa! untuk SD/MI Kelas I (A. Purwantini, Ed.; Edisi Revisi). Kementerian Pendidikan, Kebudayaan, Riset, dan Teknologi. https://buku.kemdikbud.go.id
Enrique, G., Alfina, I., & Yulianti, E. (2024). Javanese part-of-speech tagging using cross-lingual transfer learning. IAES International Journal of Artificial Intelligence, 13(3), 3498–3509. https://doi.org/10.11591/ijai.v13.i3.pp3498-3509
Kamayani, M. (2019). Perkembangan Part-of-Speech Tagger Bahasa Indonesia. Jurnal Linguistik Komputasional (JLK), 2(2), 34. https://doi.org/10.26418/jlk.v2i2.20
Kurniawan, M., Kusrini, K., & Arief, M. R. (2022). Part of Speech Tagging Pada Teks Bahasa Indonesia dengan BiLSTM + CNN + CRF dan ELMo. Jurnal Eksplora Informatika, 11(1), 29–37. https://doi.org/10.30864/eksplora.v11i1.506
Kusuma, T. B., Widiartha, I. M., Suhartana, I. K. G., & Putra, I. G. N. A. C. (2026). Implementasi Attention-Based BiLSTM dengan LoRA Parameter Tuning untuk Analisis Sentimen Ulasan Destinasi Wisata. Techno.Com, 25(1). https://doi.org/10.62411/tc.v25i1.15089
Manalu, D., Puspaningrum, E. Y., & Putra, C. A. (2026). Predictive Maintenance dengan Klasifikasi One-vs-Rest dan Random Search. CICES (Cyberpreneurship Innovative and Creative Exact and Social Science), 12(1). https://doi.org/10.33050/jqpbvj44
Nurwanda, N., Suarna, N., & Prihartono, W. (2024). Penerapan NLP (Natural Language Processing) dalam Analisis Sentimen Pengguna Telegram di PlayStore. JATI (Jurnal Mahasiswa Teknik Informatika), 8(2), 1841–1846. https://doi.org/10.36040/jati.v8i2.8469
Pinem, T. H., & Putra, Z. P. (2025). Evaluasi Kinerja Algoritma Klasifikasi Deep Learning dalam Prediksi Diabetes. Jurnal Ilmiah FIFO, 17(1), 17. https://doi.org/10.22441/fifo.2025.v17i1.003
Putra, R. R., Richasdy, D., & Ihsan, A. F. (2023). Part-of-Speech Tagging Implementation on Telkom University News using Bidirectional LSTM Method. Jurnal Media Informatika Budidarma, 7(1), 360–368. https://doi.org/10.30865/mib.v7i1.5506
Riezler, S., & Hagmann, M. (2022). Validity, Reliability, and Significance: Empirical Methods for NLP and Data Science. Dalam Synthesis Lectures on Human Language Technologies (Vol. 14, Nomor 6, hlm. 1–147). Morgan and Claypool Publishers. https://doi.org/10.2200/S01137ED1V01Y202110HLT055
Rostam, Z. R. K., & Kertész, G. (2025). Advances in Pre-trained Language Models for Domain-Specific Text Classification: A Systematic Review. Dalam ACM Transactions on Intelligent Systems and Technology (Vol. 16, Nomor 6). Association for Computing Machinery. https://doi.org/10.1145/3763002
Sari, J. R., Sadik, K., Soleh, A. M., & Suhaeni, C. (2025). Evaluasi Model Klasifikasi dalam Deteksi Penipuan Transaksi: Studi Kasus pada Dat Tidak Seimbang. Jurnal Gaussian, 14(2), 565–576. https://doi.org/10.14710/j.gauss.14.2.565-576
Simanjuntak, A., Lumbantoruan, R., Sianipar, K., Gultom, R., Simaremare, M., Situmeang, S., & Panggabean, E. (2024). Research and Analysis of IndoBERT Hyperparameter Tuning in Fake News Detection. Jurnal Nasional Teknik Elektro dan Teknologi Informasi, 13(1), 60–67. https://doi.org/10.22146/jnteti.v13i1.8532
Widhiyasana, Y., Semiawan, T., Mudzakir, I. G. A., & Noor, M. R. (2021). Penerapan Convolutional Long Short-Term Memory untuk Klasifikasi Teks Berita Bahasa Indonesia. Jurnal Nasional Teknik Elektro dan Teknologi Informasi, 10(4), 354–361. https://doi.org/10.22146/jnteti.v10i4.2438
Yazid, A. S., & Winarko, E. (2023). Fine-Tuning BERT untuk Menangani Ambiguitas Pada POS Tagging Bahasa Indonesia. Jurnal Linguistik Komputasional (JLK), 6(2), 57-64. https://doi.org/10.26418/jlk.v6i2.148
Bila bermanfaat silahkan share artikel ini
Berikan Komentar Anda terhadap artikel Analisis Performa Metode BiLSTM untuk POS Tagging pada Kalimat Bahasa Indonesia Tingkat Dasar
Copyright (c) 2026 Dhafa Widy Pratama, Ledy Elsera Astrianty
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under Creative Commons Attribution 4.0 International License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (Refer to The Effect of Open Access).
