Integrating SeqGAN and NLP Augmentation to Improve Generalization in Sentiment Classification Models

Authors

Shafin Mahmood
Department of Computer Science, American International University-Bangladesh
Sayma Akter Trina
Department of Computer Science, American International University-Bangladesh
Md. Easin Khondokar
Department of Computer Science, American International University-Bangladesh
Nuzhat Tabassum
Department of Computer Science, American International University-Bangladesh
DOI: https://doi.org/10.52549/ijeei.v14i3.7934
Download PDF
Published: 10-09-2026 • Issue: Vol. 14 No. 3: September 2026 • Section: Regular paper

Abstract

Sentiment analysis is essential for understanding people's experiences and social media trends. However, challenges such as imbalanced datasets, limited data availability, and the informal nature of user-generated content often degrade a model's ability to achieve higher accuracy. This study presents that data augmentation using SeqGAN can effectively addresses these issues and improve prediction performance compared to traditional NLP-based augmentation techniques. A variety of Machine Learning (ML), Deep Learning (DL), hybrid, and GAN-based models were evaluated to predict sentiment from social media comments. After generating synthetic data using both SeqGAN and NLP-based techniques, all models were retrained, and notable improvements in accuracy were observed with the SeqGAN-augmented dataset. The use of synthetically generated data led to a statistically significant increase in classification performance, with a mean absolute accuracy improvement of approximately 4–5% across the full dataset. These findings highlight that GAN-based data augmentation, particularly through tailored synthetic sample generation, can enhance model robustness and boost sentiment prediction accuracy more effectively than conventional augmentation approaches.
Keywords: Sentiment Analysis; GAN Augmentation; NLP Augmentation; Deep Learning; Machine Learning

References

[1] R. Sharma, S. Nigam, and R. Jain, “Polarity detection at sentence level,” Int J Comput Appl, vol. 86,

no. 11, pp. 29–33, 2014.

[2] V. Umarani, A. Julian, and J. Deepa, “Sentiment analysis using various ML and DL techniques,” J Niger

Soc Phys Sci, vol. 3, pp. 385–394, 2021.

[3] F. Jemai, M. Hayouni, and S. Baccar, “Sentiment analysis using ML algorithms,” in 2021 IEEE IWCMC,

2021, pp. 775–779.

[4] C. Shorten, T. Khoshgoftaar, and B. Furht, “Text data augmentation for DL,” J Big Data, vol. 8, 2021.

[5] P. Ghadekar, M. Jamble, A. Jaybhay, B. Jagtap, A. Joshi, and H. More, “Text data augmentation using

generative adversarial networks, back translation and EDA,” in 2023, pp. 391–401.

[6] H. Q. Abonizio, E. C. Paraiso, and S. Barbon, “Toward text data augmentation for sentiment analysis,”

IEEE Transactions on Artificial Intelligence, vol. 3, no. 5, pp. 657–668, 2022.

[7] Z. Khanam, B. Alwasel, H. Sirafi, and M. Rashid, “Fake news detection using ML approaches,” IOP Conf

Ser: Mater Sci Eng, vol. 1099, p. 012040, 2021.

[8] W. Cavnar, J. Trenkle et al., “N-gram-based text categorization,” in Proc SDAIR-94, 3rd Annu Symp Doc

Anal Inf Retr, 1994, p. 14.

[9] ˙I. Tarımer, A. C¸ oban, and A. Kocaman, “Sentiment analysis on IMDB movie comments and twitter data

by ML and vector space techniques,” arXiv preprint arXiv:1903.11983, 2019.

[10] A. McCallum, K. Nigam et al., “A comparison of event models for naive bayes text classification,” in

AAAI-98 Workshop Learn Text Categorization, 1998, pp. 41–48.

[11] J. Rennie, L. Shih, J. Teevan, and D. Karger, “Tackling the poor assumptions of naive bayes text classifiers,”

in Proc 20th Int Conf Mach Learn (ICML-03), 2003, pp. 616–623.

[12] L. Jiang, C. Li, S. Wang, and L. Zhang, “Deep feature weighting for naive bayes and its application to

text classification,” Eng Appl Artif Intell, vol. 52, pp. 26–39, 2016.

[13] C. Aggarwal and C. Zhai, “A survey of text classification algorithms,” in Mining Text Data. Springer,

2012, pp. 163–222.

[14] X. Gong, W. Ying, S. Zhong, and S. Gong, “Text sentiment analysis based on transformer and augmentation,”

Frontiers in Psychology, vol. Volume 13 - 2022, 2022.

[15] K. Alahmadi, S. Alharbi, J. Chen, and X. Wang, “Generalizing sentiment analysis: a review of progress,

challenges, and emerging directions,” Social Network Analysis and Mining, vol. 15, 04 2025.

[16] A. Igali, A. Abdrakhman, Y. Torekhan, and P. Shamoi, “Tracking emotional dynamics in chat conversations:

A hybrid approach using DistilBERT and emoji sentiment analysis,” arXiv preprint

arXiv:2408.01838, 2024.

[17] F. Masoumi and M. Bahrani, “Utilizing distilBERT transformer model for sentiment classification of

COVID-19’s persian open-text responses,” arXiv preprint arXiv:2212.08407, 2022.

[18] X.-Y. Fu, C. Chen, M. Laskar, S. Gardiner, P. Hiranandani, and S. Tn, “Entity-level sentiment analysis in

contact center telephone conversations,” arXiv preprint arXiv:2210.13401, 2022.

[19] K. Tan, C. Lee, K. Anbananthen, and K. Lim, “RoBERTa-LSTM: a hybrid model for sentiment analysis

with transformer and recurrent neural network,” IEEE Access, vol. 10, pp. 21 517–21 525, 2022.

[20] M. Cliche, “BB twtr at semeval-2017 task 4: Twitter sentiment analysis with CNNs and LSTMs,” arXiv

preprint arXiv:1704.06125, 2017.

[21] A. S. Imran, R. Yang, Z. Kastrati, S. M. Daudpota, and S. Shaikh, “The impact of synthetic text generation

for sentiment analysis using GAN based models,” Egyptian Informatics Journal, vol. 23, no. 3, pp. 547–

557, 2022. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S1110866522000342

[22] R. Yang and M. Edalati, “Using GAN-based models to sentimental analysis on imbalanced datasets in

education domain,” arXiv preprint arXiv:2108.12061, 2021.

[23] P. Sundarreson and S. Kumarapathirage, “SentiGEN: Synthetic data generator for sentiment analysis,”

Journal of Computing Theories and Applications, vol. 1, pp. 461–477, 04 2024.

[24] J. Luo, M. Bouazizi, and T. Ohtsuki, “Data augmentation for sentiment analysis using sentence

compression-based SeqGAN with data screening,” IEEE Access, vol. 9, pp. 99 922–99 931, 2021.

[25] K. Wahyu Trisna, J. Huang, Y. Chen, and I. G. J. E. Putra, “Dynamic text augmentation for robust sentiment

analysis: Enhancing model performance with EDA and multi-channel CNN,” IEEE Access, vol. 13,

pp. 31 978–31 991, 2025.

[26] P. Eigenschink, T. Reutterer, S. Vamosi, R. Vamosi, C. Sun, and K. Kalcher, “Deep generative models for

synthetic data: A survey,” IEEE Access, vol. 11, pp. 47 304–47 320, 2023.

[27] K. Silva, B. Can, R. Sarwar, F. Blain, and R. Mitkov, “Text data augmentation using generative adversarial

networks – a systematic review,” Journal of Computational and Applied Linguistics, vol. 1, pp. 6–38, 07

2023.

[28] M. Mohammadi, M. R. Amin, and S. Tavakoli, “Boosting sentiment analysis in Persian through a

GAN-based synthetic data augmentation method,” in Proceedings of the 1st Workshop on NLP for

Languages Using Arabic Script, M. El-Haj, A. Haddad, C. Amol, S. Ahmadi, H. Paterson III, I. Ezeani,

S. Ezzini, and P. Rayson, Eds. Abu Dhabi, UAE: Association for Computational Linguistics, Jan. 2025,

pp. 54–63. [Online]. Available: https://aclanthology.org/2025.abjadnlp-1.7/

[29] Kazanova, “Sentiment140 dataset with 1.6 million tweets,” https://www.kaggle.com/datasets/kazanova/

sentiment140, 2019, accessed: 2026-07-04.

[30] M. Ahmad, S. Aftab, and I. Ali, “Sentiment analysis of tweets using SVM,” Int J Comput Appl, vol. 177,

pp. 975–8887, 2017.

[31] M. Rezwanul, A. Ali, and A. Rahman, “Sentiment analysis on twitter data using KNN and SVM,” Int J

Adv Comput Sci Appl, vol. 8, 2017.

[32] M. Aufar, R. Andreswari, and D. Pramesti, “Sentiment analysis on youtube social media using decision

tree and random forest algorithm: A case study,” in 2020 International Conference on Data Science and

Its Applications (ICoDSA). IEEE, 2020, pp. 1–7.

[33] H. Parveen and S. Pandey, “Sentiment analysis on twitter data-set using naive bayes algorithm,” in 2016

2nd Int Conf Appl Theor Comput Commun Technol (iCATccT), 2016, pp. 416–419.

[34] P. S. Reddy, D. R. Sri, C. S. Reddy, and S. Shaik, “Sentimental analysis using logistic regression,” International

Journal of Engineering Research and Applications, vol. 11, no. 7, pp. 36–40, 2021.

[35] M. Umer, Z. Imtiaz, M. Ahmad, M. Nappi, C. Medaglia, G. Choi, and A. Mehmood, “Impact of convolutional

neural network and fasttext embedding on text classification,” Multimed Tools Appl, vol. 82,

2022.

[36] D. Murthy, S. Allu, B. Andhavarapu, and M. Bagadi, “Text based sentiment analysis using LSTM,” Int J

Eng Res Technol V9, 2020.

[37] J. Vimali and S. Murugan, “A text based sentiment analysis model using bi-directional LSTM networks,”

in 2021 6th Int Conf Commun Electron Syst (ICCES), 2021, pp. 1652–1658.

[38] Y. Lee, J. Saxe, and R. Harang, “CATBERT: context-aware tiny BERT for detecting social engineering

emails,” arXiv preprint arXiv:2010.03484, 2020.

[39] M. Rhanoui, M. Mikram, S. Yousfi, and S. Barzali, “A CNN-BiLSTM model for document-level sentiment

analysis,” Mach Learn Knowl Extr, vol. 1, no. 3, pp. 832–847, 2019.

[40] F. Sun and N. Chu, “Text sentiment analysis based on CNN-BiLSTM-attention model,” in 2020 International

Conference on Robots & Intelligent System (ICRIS), 2020, pp. 749–752.

[41] R. Shu, H. Bui, and S. Ermon, “Ac-gan learns a biased distribution,” in NIPS Workshop on Bayesian Deep

Learning, vol. 8, 2017, p. 34.

[42] Y. Liu, Q. Zhao, and C. Jiang, “Conditional image generation using feature-matching GAN,” in 2017 10th

Int Congr Image Signal Process, BioMedical Eng Informatics (CISP-BMEI), 2017, pp. 1–5.

[43] M. Sokolova and G. Lapalme, “A systematic analysis of performance measures for classification tasks,”

Information Processing & Management, vol. 45, no. 4, pp. 427–437, 2009.

Author Biographies

Shafin Mahmood
Shafin is a B.Sc. Computer Science student at American International University-Bangladesh. His areas of research are Artificial Intelligence, Machine Learning, Deep Learning, Natural Language Processing, generative adversarial networks, and high-accuracy AI-driven decision frameworks. He is co-author of the 2025 MDPI Drones paper "Multi-Agent Actor–Critic Frameworks for UAV Swarm Networks" (Q1, IF 4.4). He is committed to creating intelligent distributed systems and describing theoretical models as concrete, reproducible research outputs. He can be contacted at: shafin26103@gmail.com and personal website https://shafin-mahmood.github.io/
Sayma Akter Trina
Sayma is a B.Sc. student in Computer Science at the American International University-Bangladesh, with a strong academic focus on artificial intelligence, machine learning, and data-driven systems. Her research interests include machine learning, deep learning, and algorithm optimization. She is dedicated to building interpretable and scalable intelligent systems that integrate theoretical advancements with real-world applications. She can be contacted at trinasayma5191@gmail.com
Md. Easin Khondokar
Easin is a B.Sc. student in Computer Science at the American International University-Bangladesh (AIUB) with research interests in Machine Learning, Deep Learning, Natural Language Processing, and Cyber Security. His research and development focus lies in designing and implementing AI-based systems that integrate deep learning and NLP to solve real-world challenges through intelligent automation and secure computing solutions. He can be contacted at: easinkhandokarcse@gmail.com.
Nuzhat Tabassum
Nuzhat Tabassum is an Assistant Professor at the American International University-Bangladesh (AIUB), Faculty of Science and Technology, Department of Computer Science. She completed her M.Sc. in Artificial Intelligence from Warsaw University of Technology (WUT), Warsaw, Poland, and B.Sc. in Computer Science and Engineering from Chittagong University of Engineering and Technology, Bangladesh. Prior to AIUB, she worked as an Assistant Professor at the Military Institute of Science and Technology (MIST), and before that, she was a Lecturer at BRAC University. She also worked as a Junior AI Engineer at Smartlife Robotics, Warsaw, Poland, and as a BI Specialist at Robi Axiata Limited. Nuzhat has success and experience in both industry and academia. She is committed to developing her abilities and making a significant impact in her profession. Her field of interest includes Deep learning, Machine learning, Natural Language Processing, Data Mining, Image Processing, and Augmented Reality.

How to Cite

[1]
S. Mahmood, S. A. Trina, M. E. Khondokar, and N. Tabassum, “Integrating SeqGAN and NLP Augmentation to Improve Generalization in Sentiment Classification Models”, IJEEI, vol. 14, no. 3, pp. 740–755, Sep. 2026, doi: 10.52549/ijeei.v14i3.7934.