Conference Paper2023Cited by 6

Investigation the Effectiveness of Graph-based Algorithm for Bangla Text Classification

F Dehan, M Fahim, A A Ali, MA Amin, AKMM Rahman

1st Workshop on Bangla Language Processing in conjunction with EMNLP

Association of Computational Linguistics, pp. 104-116

DOI: 10.18653/v1/2023.banglalp-1.12Open AccessURL

Abstract

In this study, we examine and analyze the behavior of several graph-based models for Bangla text classification tasks. Graph-based algorithms create heterogeneous graphs from text data. Each node represents either a word or a document, and each edge indicates relationship between any two words or word and document. We applied the BERT model and different graph-based models including TextGCN, GAT, BertGAT, and BertGCN on five different datasets including SentNoB, Sarcasm detection, BanFakeNews, Hate speech detection, and Emotion detection datasets for Bangla text. BERT’s model bested the TextGCN and the GAT models by a large difference in terms of accuracy, Macro F1 score, and weighted F1 score. BertGCN and BertGAT are shown to outperform standalone graph models and BERT model. BertGAT excelled in the Emotion detection dataset and achieved a 1%-2% performance boost in Sarcasm detection, Hate speech detection, and BanFakeNews datasets from BERT’s performance. Whereas, BertGCN outperformed BertGAT by 1% for SetNoB, and BanFakeNews datasets while beating BertGAT by 2% for Sarcasm detection, Hate Speech, and Emotion detection datasets. We also examined different variations in graph structure and analyzed their effects.

Keywords

  • hate speech detection
  • Emotion recognition
  • fake news detection
  • GAT
  • sarcasm detection
  • graph-based modeling
  • BERT
  • heterogeneous graphs
  • graph structure

References

  1. 1.Kefallinos, Dionysios, Alexandris, Georgios, Maras, Alexis, Chaidos, Panagiotis, Gomony, Manil Dev, Corporaal, Henk, Soudris, Dimitrios, Xydis, Sotirios. (2018). AI-Assisted Pipeline for Dynamic Generation of Trustworthy Health Supplement Content at Scale. DROPS (Schloss Dagstuhl – Leibniz Center for Informatics)[10.4230/lipics.cosit.2022.18]
  2. 2.Chen, Tianyu, Wang, Zeyu, Li, Lin, Li, Ding, Li, Zongyang, Chang, Xiaoning, Bian, Pan, Liang, Guangtai, Wang, Qianxiang, Xie, Tao. (2025). Detecting Functionality-Specific Vulnerabilities via Retrieving Individual Functionality-Equivalent APIs in Open-Source Repositories. [10.4230/lipics.ecoop.2025.6]
  3. 3.Yoon Kim. (2014). Convolutional Neural Networks for Sentence Classification. , 1746–1751[10.3115/v1/d14-1181]
  4. 4.Alex Sherstinsky. (2020). Fundamentals of Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) network. Physica D Nonlinear Phenomena, 404, 132306[10.1016/j.physd.2019.132306]
  5. 5.Silva, Gabriel, Rodrigues, Mário, Teixeira, António, Amorim, Marlene. (2024). Targeted Branching for the Maximum Independent Set Problem Using Graph Neural Networks. DROPS (Schloss Dagstuhl – Leibniz Center for Informatics)[10.4230/lipics.sea.2024.20]
  6. 6.Liang Yao, Chengsheng Mao, Yuan Luo. (2019). Graph Convolutional Networks for Text Classification. Proceedings of the AAAI Conference on Artificial Intelligence, 33(01), 7370–7377[10.1609/aaai.v33i01.33017370]
  7. 7.Bin Liang, Hang Su, Lin Gui, Erik Cambria, Ruifeng Xu. (2021). Aspect-based sentiment analysis via affective knowledge enhanced graph convolutional networks. Knowledge-Based Systems, 235, 107643[10.1016/j.knosys.2021.107643]
  8. 8.Abhik Bhattacharjee, Tahmid Hasan, Wasi Ahmad, Kazi Samin Mubasshir, Md Saiful Islam, Anindya Iqbal, M. Sohel Rahman, Rifat Shahriyar. (2022). BanglaBERT: Language Model Pretraining and Benchmarks for Low-Resource Language Understanding Evaluation in Bangla. Findings of the Association for Computational Linguistics: NAACL 2022, 1318–1327[10.18653/v1/2022.findings-naacl.98]
  9. 9.Nafis Irtiza Tripto, Mohammed Eunus Ali. (2018). Detecting Multilabel Sentiment and Emotions from Bangla YouTube Comments. , 1–6[10.1109/icbslp.2018.8554875]
  10. 10.Karim, M.R., Chakravarthi, B.R., McCrae, J.P., Cochez, M.. (2020). Classification Benchmarks for Under-resourced Bengali Language based on Multichannel Convolutional-LSTM Network. Publikationsdatenbank der Fraunhofer-Gesellschaft (Fraunhofer-Gesellschaft)[10.1109/dsaa49011.2020.00053]
  11. 11.Abdhullah-Al-Mamun, Shahin Akhter. (2018). Social media bullying detection using machine learning on Bangla text. , 385–388[10.1109/icece.2018.8636797]
  12. 12.Khondoker Ittehadul Islam, Sudipta Kar, Md Saiful Islam, Mohammad Ruhul Amin. (2021). SentNoB: A Dataset for Analysing Sentiment on Noisy Bangla Texts. , 3265–3271[10.18653/v1/2021.findings-emnlp.278]
  13. 13.Yazhou Zhang, Dan Ma, Prayag Tiwari, Chen Zhang, Mehedi Masud, Mohammad Shorfuzzaman, Dawei Song. (2022). Stance-level Sarcasm Detection with BERT and Stance-centered Graph Attention Networks. ACM Transactions on Internet Technology, 23(2), 1–21[10.1145/3533430]
  14. 14.Abu Nowshed Chy, Md. Hanif Seddiqui, Sowmitra Das. (2014). Bangla news classification using naive Bayes classifier. , 366–371[10.1109/iccitechn.2014.6997369]
  15. 15.Md Mahbubur Rahman, Md. Aktaruzzaman Pramanik, Rifat Sadik, Monikrishna Roy, Partha Chakraborty. (2020). Bangla Documents Classification using Transformer Based Deep Learning Models. , 1–5[10.1109/sti50764.2020.9350394]
  16. 16.Md Tanvir Alam, Md Mofijul Islam. (2018). BARD: Bangla Article Classification Using a New Comprehensive Dataset. , 1–5[10.1109/icbslp.2018.8554382]
  17. 17.Xujuan Zhou, Raj Gururajan, Yuefeng Li, Revathi Venkataraman, Xiaohui Tao, Ghazal Bargshady, Prabal D. Barua, Srinivas Kondalsamy-Chennakesavan. (2020). A survey on text classification and its applications. Web Intelligence, 18(3), 205–216[10.3233/web-200442]
  18. 18.Saifur Rahman, Partha Chakraborty. (2021). Bangla Document Classification Using Deep Recurrent Neural Network with BiLSTM. Algorithms for intelligent systems, 507–519[10.1007/978-981-33-4087-9_43]
  19. 19.Lingfei Wu, Yu Chen, Kai Shen, Xiaojie Guo, Hanning Gao, Shucheng Li, Jian Pei, Bo Long. (2023). Graph Neural Networks for Natural Language Processing: A Survey. [10.1561/9781638281436]
  20. 20.Ankita Dhar, Niladri Sekhar Dash, Kaushik Roy. (2018). Classification of Bangla Text Documents based on Inverse Class Frequency. , 1–6[10.1109/iot-siu.2018.8519866]
  21. 21.Mostaq Ahmed, Partha Chakraborty, Tanupriya Choudhury. (2021). Bangla Document Categorization Using Deep RNN Model with Attention Mechanism. Lecture notes in networks and systems, 137–147[10.1007/978-981-16-4284-5_13]
  22. 22.Abu Kowshir Bitto, Md. Hasan Imam Bijoy, Md. Shohel Arman, Imran Mahmud, Aka Das, Joy Majumder. (2023). Sentiment analysis from Bangladeshi food delivery startup based on user reviews using machine learning and deep learning. Bulletin of Electrical Engineering and Informatics, 12(4), 2282–2291[10.11591/eei.v12i4.4135]
  23. 23.Phuc Do, Trung Phan, Hung Le, Brij B. Gupta. (2020). Building a knowledge graph by using cross-lingual transfer method and distributed MinIE algorithm on apache spark. Neural Computing and Applications, 34(11), 8393–8409[10.1007/s00521-020-05495-1]
  24. 24.Niloy Farhan, Ishrat Tasnim Awishi, Md Humaion Kabir Mehedi, MD. Mustakin Alam, Annajiat Alim Rasel. (2023). Ensemble of Gated Recurrent Unit and Convolutional Neural Network for Sarcasm Detection in Bangla. , 0624–0629[10.1109/ccwc57344.2023.10099157]
  25. 25.Mehzabin Sadat Aothoi, Samin Ahsan, Najeefa Nikhat Choudhury, Annajiat Alim Rasel. (2023). Supervised Hybrid Model for Rumor Classification: A Comparative Study of Machine and Deep Learning Approaches. Lecture notes in computer science, 281–286[10.1007/978-3-031-39831-5_25]
  26. 26.Veli\v{c}kovi\'c, Petar, Cucurull, Guillem, Casanova, Arantxa, Romero, Adriana, Li\`o, Pietro, Bengio, Yoshua. (2017). Graph Attention Networks. arXiv (Cornell University)[link]
  27. 27.Thekumparampil, Kiran K., Wang, Chong, Oh, Sewoong, Li, Li-Jia. (2018). Attention-based Graph Neural Network for Semi-supervised Learning. arXiv (Cornell University)[10.48550/arxiv.1803.03735]
  28. 28.Hossain, Md Zobaer, Rahman, Md Ashraful, Islam, Md Saiful, Kar, Sudipta. (2020). BanFakeNews: A Dataset for Detecting Fake News in Bangla. arXiv (Cornell University)[10.48550/arxiv.2004.08789]
  29. 29.Lin, Yuxiao, Meng, Yuxian, Sun, Xiaofei, Han, Qinghong, Kuang, Kun, Li, Jiwei, Wu, Fei. (2021). BertGCN: Transductive Text Classification by Combining GCN and BERT. arXiv (Cornell University)[10.48550/arxiv.2105.05727]
  30. 30.Iosr Journals, Ms. Priyanka Patel1 , Ms. Khushali Mistry. (2015). A Review: Text Classification on Social Media Data. Figshare[10.6084/m9.figshare.1335951.v1]
  31. 31.Han, Soyeon Caren, Yuan, Zihan, Wang, Kunze, Long, Siqu, Poon, Josiah. (2022). Understanding Graph Convolutional Networks for Text Classification. arXiv (Cornell University)[10.48550/arxiv.2203.16060]
  32. 32.Wang, Kunze, Ding, Yihao, Han, Soyeon Caren. (2023). Graph Neural Networks for Text Classification: A Survey. arXiv (Cornell University)[10.48550/arxiv.2304.11534]

Cited By

  1. 1.Sabrim Mahmud, Samaun Rezvi Payel, Jawad Hossain, Nuzhat Tabassum Progga, Asma Mariam. (2025). Identifying Bangla Political Hate Speech using Transformer-based Approach. International Conference on Next-Generation Computing, IoT and Machine Learning (NCIM), 1–6[10.1109/ncim65934.2025.11159848]
  2. 2.Farhan Noor Dehan, Md Fahim, A. K. M. Mahabubur Rahman, M. Ashraful Amin, Amin Ahsan Ali. (2024). TinyLLM Efficacy in Low-Resource Language: An Experiment on Bangla Text Classification Task. Lecture notes in computer science, 472–487[10.1007/978-3-031-78495-8_30]
  3. 3.Md. Siam Ansary. (2025). A Novel Dataset for Identifying AI Generated Bengali Texts. , 1–4[10.1109/qpain66474.2025.11171930]
  4. 4.Md. Siam Ansary. (2026). Detection of Bangla AI ‐Generated Texts Using a Quantum Kernel Support Vector Machine. Applied AI Letters, 7(3)[10.1002/ail2.70051]
  5. 5.Aditi Nag, Sagar Singh, Anurag Singh Rathore. (2026). AI in Smart Tourism Cities. Advances in computational intelligence and robotics book series, 347–390[10.4018/979-8-3373-9700-9.ch014]
  6. 6.Sabrim Mahmud, Samaun Rezvi Payel, Jawad Hossain. (2026). PoliticaNet: A Transformer-Driven Framework for Fine-Grained Classification of Bangla Political Hate Speech in Social Media. Communications in computer and information science, 93–111[10.1007/978-3-032-18319-4_6]
← All publications