- All Journals
- Al-Qadisiyah University
- Basrah University
- Kirkuk University
- Mosul University
- The Iraqi Borad for Medical Specialization
- University of Anbar
- University of Technology
- Al Mustafa University
- Al-Zahraa University for Women
- Alnoor University
- Kunooz University
- Misan University
- Ninevah University
- Tikrit University
- University of Al-Hamdaniya
- University of Kerbala
A Comparative Study of Machine Learning Models for Persian Loanword Identification in Arabic | ||
| Al-Zahraa Journal for Artificial Intelligence Applications | ||
| Article 2, Volume 0, Issue 1, September 2026, Pages 14-23 PDF (665.48 K) | ||
| Document Type: Research Article | ||
| DOI: 10.66723/zjaia.2026.164425.1004 | ||
| Authors | ||
| Radhwan Adnan Dakhil* 1; Murtaja S. Jalawkhan2 | ||
| 1College of Information Technology Engineering, Al-Zahraa University for Women, Karbala, Iraq | ||
| 2College of Computer Science and Information Technology, University of Kerbala, Karbala, Iraq | ||
| Abstract | ||
| The phenomenon of word borrowing between the Arabic and Persian languages represents a rich field for historical linguistic studies, yet it lacks systematic computational studies aimed at automating the classification process. The main challenge in this domain is the inherent problem of data imbalance, where the number of native words significantly surpasses their borrowed counterparts. This research presents the first benchmark study to evaluate the performance of classical machine learning models on the task of classifying Persian loanwords in Arabic, with a focus on handling data imbalance. Using a purpose-built dataset of 5,258 words (936 Persian, 4,322 non-Persian), a series of systematic experiments were conducted. The experiments began by evaluating baseline models, followed by the application of class weighting, and finally, the use of the Synthetic Minority Over-sampling Technique (SMOTE). The results showed that the Logistic Regression model with the class_weight parameter set to balanced achieved the best performance, reaching a recall of 70% for the Persian class (the minority class), significantly outperforming other models. This research not only provides the first reproducible baseline for this linguistic problem but also presents a deep linguistic analysis of the model's errors, proving that these errors are not random but rather reveal complex linguistic phenomena related to nativization, derivation, and the historical analysis of words. While this method proved highly appropriate for tabular feature extraction and baseline evaluation, it is not the most advanced technique available. The study acknowledges the lack of Deep Neural Networks (DNNs) as a limitation. | ||
| Keywords | ||
| Loanword Identification; Computational Linguistics; Natural Language Processing; Arabic; Persian; Imbalanced Data; Machine Learning; Benchmark | ||
| References | ||
|
[1] A. Khrisat, "Language's borrowings: The role of the borrowed and Arabized words in enriching Arabic language," American Journal of Humanities and Social Sciences, 2014, doi: 10.11634/232907811402533.
[2] A. K. Jassim, M. J. Hamzah, and A. H. Aliwy, "Using Graph Mining Method in Analyzing Turkish Loanwords Derived from Arabic Language," Baghdad Science Journal, vol. 19, no. 6, p. 4, 2022, doi: 10.21123/bsj.2022.6008.
[3] C. Mi, Y. Yang, L. Wang, X. Zhou, and T. Jiang, "A neural network based model for loanword identification in Uyghur," in Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018), 2018, doi: 10.1109/NLP.2018.8540775.
[4] C. Mi, Y. Yang, X. Zhou, L. Wang, X. Li, and T. Jiang, "Recurrent neural network based loanwords identification in Uyghur," in Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, 2016: Waseda University, pp. 209-217.
[5] C. Mi, Y. Yang, L. Wang, X. Zhou, and T. Jiang, "Toward better loanword identification in Uyghur using cross-lingual word embeddings," in Proceedings of the 27th International Conference on Computational Linguistics, 2018, pp. 3027-3037.
[6] C. Mi, "Improving the Robustness of Loanword Identification in Social Media Texts," ACM Transactions on Asian and Low-Resource Language Information Processing, vol. 22, no. 4, pp. 1-19, 2023, doi: 10.1145/3572773.
[7] C. Mi, S. Zhu, and R. Nie, "Improving Loanword Identification in Low‐Resource Language with Data Augmentation and Multiple Feature Fusion," Computational Intelligence and Neuroscience, vol. 2021, no. 1, p. 9975078, 2021, doi: 10.1155/2021/9975078.
[8] C. Mi, L. Xie, and Y. Zhang, "Loanword identification in low-resource languages with minimal supervision," ACM Transactions on Asian and Low-Resource Language Information Processing (TALLIP), vol. 19, no. 3, pp. 1-22, 2020, doi: 10.1145/3374212.
[9] J. E. Miller, T. Tresoldi, R. Zariquiey, C. A. Beltran Castanon, N. Morozova, and J.-M. List, "Using lexical language models to detect borrowings in monolingual wordlists," Plos one, vol. 15, no. 12, p. e0242709, 2020, doi: 10.1371/journal.pone.0242709.
[10] A. Nath, S. M. Saravani, I. Khebour, S. Mannan, Z. Li, and N. Krishnaswamy, "A generalized method for automated multilingual loanword detection," in Proceedings of the 29th international conference on computational linguistics, 2022, pp. 4996-5013.
[11] C. Fourrier and B. Sagot, "Comparing statistical and neural models for learning sound correspondences," in LT4HALA 2020-First Workshop on Language Technologies for Historical and Ancient Languages, 2020.
[12] Y. Tsvetkov and C. Dyer, "Cross-lingual bridges with models of lexical borrowing," Journal of Artificial Intelligence Research, vol. 55, pp. 63-93, 2016, doi: 10.1613/jair.4786.
[13] T. Ylonen. "Arabic machine-readable dictionary." kaikki.org. https://kaikki.org/dictionary/Arabic/index.html (accessed.
[14] K. Sparck Jones, "A statistical interpretation of term specificity and its application in retrieval," Journal of documentation, vol. 28, no. 1, pp. 11-21, 1972, doi: 10.1108/eb026526.
[15] G. H. John and P. Langley, "Estimating continuous distributions in Bayesian classifiers," arXiv preprint arXiv:1302.4964, 2013, doi: 10.48550/arXiv.1302.4964.
[16] D. W. Hosmer Jr, S. Lemeshow, and R. X. Sturdivant, Applied logistic regression. John Wiley & Sons, 2013, doi: 10.1002/9781118548387.
[17] C. Cortes and V. Vapnik, "Support-vector networks," Machine learning, vol. 20, no. 3, pp. 273-297, 1995, doi: 10.1007/BF00994018.
[18] R. Genuer and J.-M. Poggi, "Random forests," in Random forests with R: Springer, 2020, pp. 33-55, doi: 10.1023/A:1010933404324.
[19] N. V. Chawla, K. W. Bowyer, L. O. Hall, and W. P. Kegelmeyer, "SMOTE: synthetic minority over-sampling technique," Journal of artificial intelligence research, vol. 16, pp. 321-357, 2002, doi: 10.1613/jair.953.
[20] N. Gadelii, "The morphological integration of loanwords into Modern Standard Arabic," 2015.
[21] U. Tadmor, "Loanwords in the world’s languages: Findings and results," Loanwords in the world’s languages: A comparative handbook, vol. 55, p. 75, 2009, doi: 10.1515/9783110218442.
[22] M. Cennamo, "Aspects of grammaticalization and reanalysis in the voice domain in the transition from Latin to early Italo-Romance," in Perspectives on Language structure and language change: John Benjamins Publishing Company, 2019, pp. 205-231, doi: 10.1075/cilt.345.09cen.
[23] T. Prasad, A course in linguistics. PHI Learning Pvt. Ltd., 2019, doi: 10.1111/j.1467-1770.1958.tb00870.x.
[24] L. Zhang, R. Fabri, J. Nerbonne, and J. Nerbonne, "Detecting loan words computationally," in Variation rolls the dice: A worldwide collage in honour of Salikoko S. Mufwene: John Benjamins Publishing Company, 2021, pp. 269-288, doi: 10.1075/coll.59.11zha.
[25] R. Lass, Historical linguistics and language change. Cambridge University Press, 1997, doi: 10.1017/CBO9780511620928.
| ||
|
Statistics Article View: 42 PDF Download: 10 |
||