TurkRetNet: A Novel Retentive Network for Turkish Text Classification

Jahanggir Hossain Setu; Syed Tangim Pasha; Ashraful Islam; Nabarun Halder; Asif Mahmud; M Ashraful Amin

doi:10.1109/scored68498.2025.11399052

Conference Paper2025

TurkRetNet: A Novel Retentive Network for Turkish Text Classification

Jahanggir Hossain Setu, Syed Tangim Pasha, Ashraful Islam, Nabarun Halder, Asif Mahmud, M Ashraful Amin

2025 IEEE 23rd Student Conference on Research and Development (SCOReD)

In: 2025 IEEE 23rd Student Conference on Research and Development (SCOReD)

IEEE, pp. 1-6

DOI: 10.1109/scored68498.2025.11399052 URL

Abstract

In the era of rapidly growing digital information, the classification of unstructured data from sources, e.g., Internet news articles, has become increasingly challenging. This study introduces the Turkish Retentive Network (TurkRetNet), a novel neural network architecture implemented to improve text classification tasks specifically for the Turkish language, a low-resource language with limited Natural Language Processing (NLP) resources. To address the common issue of class imbalance in real-world datasets, this study also implemented the KeNet data augmentation technique. The performance results showed that TurkRetNet considerably outperformed traditional models, e.g., Bidirectional Encoder Representations from Transformer for Turkish (BERTurk), and Convolutional BERT for Turkish (ConvBERTurk). The balanced dataset resulted in substantial improvements over the imbalanced dataset, with classification accuracy increasing by 6.06%, macro precision by 7.89%, macro recall by 9.52%, and macro F1-score by 8.86%. This research emphasises the efficacy of combining advanced neural architectures with effective data augmentation approaches that specifically target the distinct issues presented by low-resource languages. The combined application of TurkRetNet and KeNet set a new benchmark in Turkish text classification tasks. This study contributes to the advancement of NLP for under-resourced languages.

Keywords

TurkishConvolutional neural networkEncoderMacroBenchmark (surveying)Data setArtificial neural networkTransformerDeep learning

CCDS Authors

Ashraful Islam, PhD

Center Director, CCDS
Director, HCI Wing

Assistant Professor

Department of Computer Science and Engineering

Independent University, Bangladesh (IUB)

Research Interests

Human-Computer Interaction, AI for Social Good, AI for Public Health, AI for Impact

References

1.Jakubův, Jan, Chvalovský, Karel, Goertzel, Zarathustra, Kaliszyk, Cezary, Olšák, Mirek, Piotrowski, Bartosz, Schulz, Stephan, Suda, Martin, Urban, Josef. (2023). Exploiting Generative AI to Scale up Intelligent Tutoring Systems. DROPS (Schloss Dagstuhl – Leibniz Center for Informatics)[10.4230/lipics.itp.2023.19]
2.Nitin Indurkhya, Fred J. Damerau. (2010). Handbook of Natural Language Processing. [10.1201/9781420085938]
3.Alper Kursat Uysal, Serkan Gunal. (2013). The impact of preprocessing on text classification. Information Processing & Management, 50(1), 104–112[10.1016/j.ipm.2013.08.006]
4.Hasibe Busra Dogru, Sahra Tilki, Akhtar Jamil, Alaa Ali Hameed. (2021). Deep Learning-Based Classification of News Texts Using Doc2Vec Model. , 91–96[10.1109/caida51941.2021.9425290]
5.Pinar Savci, Bihter Das. (2023). Comparison of pre-trained language models in terms of carbon emissions, time and accuracy in multi-label text classification using AutoML. Heliyon, 9(5), e15670[10.1016/j.heliyon.2023.e15670]
6.Engin Demir, Metin Bilgin. (2023). Sentiment Analysis from Turkish News Texts with BERT-Based Language Models and Machine Learning Algorithms. , 01–04[10.1109/ubmk59864.2023.10286719]
7.Ashok Basnet, Arun K. Timalsina. (2018). Improving Nepali News Recommendation Using Classification Based on LSTM Recurrent Neural Networks. Journal of Science and Technology, 5(1), 138–142[10.1109/cccs.2018.8586815]
8.Meltem Tokgoz, Fatmanur Turhan, Necva Bolucu, Burcu Can. (2021). Tuning Language Representation Models for Classification of Turkish News. 2021 International Symposium on Electrical, Electronics and Information Engineering, 402–407[10.1145/3459104.3459170]
9.Ozge Bakay, Ozlem Ergelen, Olcay Taner Yildiz. (2019). Integrating Turkish Wordnet KeNet to Princeton WordNet: The Case of One-to-Many Correspondences. 2019 Innovations in Intelligent Systems and Applications Conference (ASYU), 1–5[10.1109/asyu48272.2019.8946386]
10.N. Disayiram, R. A. H. M. Rupasingha. (2022). A Comparative Study of Classifying English News Articles Using Machine Learning Algorithms. , 50–55[10.1109/teeccon54414.2022.9854832]
11.Ahmet Birim, Mustafa Erden, Levent M. Arslan. (2021). Zero-Shot Turkish Text Classification. , 1–4[10.1109/siu53274.2021.9477864]
12.Halil Ibrahim Okur, Ahmet Sertbas. (2021). Pretrained Neural Models for Turkish Text Classification. 2021 6th International Conference on Computer Science and Engineering (UBMK), 174–179[10.1109/ubmk52708.2021.9558878]
13.Avani Sakhapara, Dipti Pawade, Hardik Chapanera, Harshal Jani, Darpan Ramgaonkar. (2018). Segregation of Similar and Dissimilar Live RSS News Feeds Based on Similarity Measures. Advances in intelligent systems and computing, 333–344[10.1007/978-981-13-1274-8_26]
14.Zekeriya Anil Guven. (2024). Performance Comparison of Large Language Models, GPT and Gemini on Turkish News Classification Task. Research Square[10.21203/rs.3.rs-4351735/v1]

← All publications

Abstract

References

1.Jakubův, Jan, Chvalovský, Karel, Goertzel, Zarathustra, Kaliszyk, Cezary, Olšák, Mirek, Piotrowski, Bartosz, Schulz, Stephan, Suda, Martin, Urban, Josef. (2023). Exploiting Generative AI to Scale up Intelligent Tutoring Systems. DROPS (Schloss Dagstuhl – Leibniz Center for Informatics)[10.4230/lipics.itp.2023.19]

2.Nitin Indurkhya, Fred J. Damerau. (2010). Handbook of Natural Language Processing. [10.1201/9781420085938]

3.Alper Kursat Uysal, Serkan Gunal. (2013). The impact of preprocessing on text classification. Information Processing & Management, 50(1), 104–112[10.1016/j.ipm.2013.08.006]

4.Hasibe Busra Dogru, Sahra Tilki, Akhtar Jamil, Alaa Ali Hameed. (2021). Deep Learning-Based Classification of News Texts Using Doc2Vec Model. , 91–96[10.1109/caida51941.2021.9425290]

5.Pinar Savci, Bihter Das. (2023). Comparison of pre-trained language models in terms of carbon emissions, time and accuracy in multi-label text classification using AutoML. Heliyon, 9(5), e15670[10.1016/j.heliyon.2023.e15670]

6.Engin Demir, Metin Bilgin. (2023). Sentiment Analysis from Turkish News Texts with BERT-Based Language Models and Machine Learning Algorithms. , 01–04[10.1109/ubmk59864.2023.10286719]

7.Ashok Basnet, Arun K. Timalsina. (2018). Improving Nepali News Recommendation Using Classification Based on LSTM Recurrent Neural Networks. Journal of Science and Technology, 5(1), 138–142[10.1109/cccs.2018.8586815]

8.Meltem Tokgoz, Fatmanur Turhan, Necva Bolucu, Burcu Can. (2021). Tuning Language Representation Models for Classification of Turkish News. 2021 International Symposium on Electrical, Electronics and Information Engineering, 402–407[10.1145/3459104.3459170]

9.Ozge Bakay, Ozlem Ergelen, Olcay Taner Yildiz. (2019). Integrating Turkish Wordnet KeNet to Princeton WordNet: The Case of One-to-Many Correspondences. 2019 Innovations in Intelligent Systems and Applications Conference (ASYU), 1–5[10.1109/asyu48272.2019.8946386]

10.N. Disayiram, R. A. H. M. Rupasingha. (2022). A Comparative Study of Classifying English News Articles Using Machine Learning Algorithms. , 50–55[10.1109/teeccon54414.2022.9854832]

11.Ahmet Birim, Mustafa Erden, Levent M. Arslan. (2021). Zero-Shot Turkish Text Classification. , 1–4[10.1109/siu53274.2021.9477864]

12.Halil Ibrahim Okur, Ahmet Sertbas. (2021). Pretrained Neural Models for Turkish Text Classification. 2021 6th International Conference on Computer Science and Engineering (UBMK), 174–179[10.1109/ubmk52708.2021.9558878]

13.Avani Sakhapara, Dipti Pawade, Hardik Chapanera, Harshal Jani, Darpan Ramgaonkar. (2018). Segregation of Similar and Dissimilar Live RSS News Feeds Based on Similarity Measures. Advances in intelligent systems and computing, 333–344[10.1007/978-981-13-1274-8_26]

14.Zekeriya Anil Guven. (2024). Performance Comparison of Large Language Models, GPT and Gemini on Turkish News Classification Task. Research Square[10.21203/rs.3.rs-4351735/v1]