Perbandingan Metode Vektorisasi Teks dan Algoritma Pembelajaran Mesin Klasik untuk Klasifikasi Teks Bahasa Arab

Haidhi Angkawijana Tedja

doi:10.51519/journalcisa.v7i3.793

Authors

Haidhi Angkawijana Tedja Telkom Corporate University

DOI:

10.51519/journalcisa.v7i3.793

Abstract

Abstract

Klasifikasi teks bahasa Arab merupakan tugas Natural Language Processing (NLP) yang menantang karena morfologi bahasa Arab yang kaya, penggunaan harakat, serta keberagaman antara Modern Standard Arabic (MSA) dan ragam dialek. Penelitian ini membandingkan tiga metode vektorisasi teks — Term Frequency–Inverse Document Frequency (TF-IDF), Count Vectorizer (CV), dan Word2Vec — yang dikombinasikan dengan empat algoritma pembelajaran mesin klasik: Logistic Regression, Linear Support Vector Machine (SVM), Decision Tree, dan Multinomial Naïve Bayes. Eksperimen dilakukan pada tiga dataset publik dengan karakteristik berbeda, yaitu SANAD (berita, 7 kelas), HARD (ulasan hotel, 4 kelas), dan OCLAR (ulasan dialek Lebanon, 5 kelas). Pra-pemrosesan meliputi pembersihan teks, penghapusan harakat, stemming ISRI, dan penghapusan stop words. Hasil menunjukkan metode berbasis frekuensi (TF-IDF dan CV) secara konsisten mengungguli Word2Vec pada model klasik, dengan akurasi terbaik 97,5% pada SANAD (SVM + TF-IDF), 70,7% pada HARD (Logistic Regression + TF-IDF), dan 62,1% pada OCLAR (Logistic Regression + CV). Decision Tree berkinerja terendah akibat curse of dimensionality, sedangkan Word2Vec dengan agregasi average hanya kompetitif pada dataset besar. Temuan ini menegaskan representasi berbasis frekuensi tetap menjadi baseline yang kuat dan efisien untuk klasifikasi teks bahasa Arab menggunakan model pembelajaran mesin klasik.

References

O. Einea, A. Elnagar, and R. Al Debsi, “SANAD: Single-label Arabic news articles dataset for automatic text categorization,” Data in Brief, vol. 25, art. 104076, 2019, doi: 10.1016/j.dib.2019.104076.

S. F. Sabbeh and H. A. Fasihuddin, “A comparative analysis of word embedding and deep learning for Arabic sentiment classification,” Electronics, vol. 12, no. 6, art. 1425, 2023, doi: 10.3390/electronics12061425.

F. Assiri and N. Alghamdi, “A comparative study of word embedding models for Arabic text,” IJCSNS International Journal of Computer Science and Network Security, vol. 22, no. 8, pp. 399–406, 2022.

A. Elnagar, R. Al-Debsi, and O. Einea, “Arabic text classification using deep learning models,” Information Processing & Management, vol. 57, no. 1, art. 102121, 2020, doi: 10.1016/j.ipm.2019.102121.

G. Salton and C. Buckley, “Term-weighting approaches in automatic text retrieval,” Information Processing & Management, vol. 24, no. 5, pp. 513–523, 1988, doi: 10.1016/0306-4573(88)90021-0.

T. Mikolov, K. Chen, G. Corrado, and J. Dean, “Efficient estimation of word representations in vector space,” arXiv preprint arXiv:1301.3781, 2013, doi: 10.48550/arXiv.1301.3781.

A. Al-Qerem, M. Raja, S. Taqatqa, and M. R. A. Sara, “Utilizing deep learning models (RNN, LSTM, CNN-LSTM, and Bi-LSTM) for Arabic text classification,” in Artificial Intelligence-Augmented Digital Twins, Studies in Systems, Decision and Control, vol. 503. Cham, Switzerland: Springer, 2024, pp. 267–278, doi: 10.1007/978-3-031-43490-7_22.

W. Alosaimi et al., “ArabBert-LSTM: Improving Arabic sentiment analysis based on transformer model and Long Short-Term Memory,” Frontiers in Artificial Intelligence, vol. 7, art. 1408845, 2024, doi: 10.3389/frai.2024.1408845.

H. Alamoudi et al., “Arabic sentiment analysis for student evaluation using machine learning and the AraBERT transformer,” Engineering, Technology & Applied Science Research, vol. 13, no. 5, pp. 11945–11952, 2023, doi: 10.48084/etasr.6347.

A. Elnagar, Y. S. Khalifa, and A. Einea, “Hotel Arabic-reviews dataset construction for sentiment analysis applications,” in Intelligent Natural Language Processing: Trends and Applications, Studies in Computational Intelligence, vol. 740. Cham, Switzerland: Springer, 2018, pp. 35–52, doi: 10.1007/978-3-319-67056-0_3.

M. Al Omari, “Opinion Corpus for Lebanese Arabic Reviews (OCLAR),” UCI Machine Learning Repository, 2019, doi: 10.24432/C52894.

O. Obeid et al., “CAMeL Tools: An open source Python toolkit for Arabic natural language processing,” in Proc. 12th Language Resources and Evaluation Conf. (LREC), Marseille, France, 2020, pp. 7022–7032.

K. Taghva, R. Elkhoury, and J. Coombs, “Arabic stemming without a root dictionary,” in Proc. Int. Conf. Information Technology: Coding and Computing (ITCC), vol. 1, Las Vegas, NV, USA, 2005, pp. 152–157, doi: 10.1109/ITCC.2005.90.

R. ?eh??ek and P. Sojka, “Software framework for topic modelling with large corpora,” in Proc. LREC 2010 Workshop on New Challenges for NLP Frameworks, Valletta, Malta, 2010, pp. 45–50.

F. Pedregosa et al., “Scikit-learn: Machine learning in Python,” Journal of Machine Learning Research, vol. 12, pp. 2825–2830, 2011.