A Novel Dual-Order Hybrid BiLSTM–BiGRU Ensemble Model for English-Bengali Code-Mixed Sentiment Analysis
Received: 6 January 2026 | Revised: 23 February 2026 and 8 March 2026 | Accepted: 9 March 2026 | Online: 6 June 2026
Corresponding author: Dalia Barua
Abstract
Sentiment analysis is a crucial task in Natural Language Processing (NLP), dealing with the identification and classification of opinionated expressions in a text. Sentiment analysis is widely used to analyze online product reviews and gain insight into customer opinions. However, the performance of sentiment analysis tasks deteriorates when code-mixing occurs, which involves using two or more languages mixed in a sentence. This type of code-mixing often occurs in Bangladeshi e-commerce product reviews. This study presents a novel Dual-Order Hybrid Bidirectional Long Short-Term Memory–Bidirectional Gated Recurrent Unit (BiLSTM–BiGRU)-based Support Vector Machine (SVM) ensemble model for English-Bangla code-mixed sentiment analysis. The proposed approach is capable of effectively capturing long-term context and patterns in code-mixed text. Experiments were conducted on both synthetic and real-world datasets and enhanced by a class-balanced distribution. The proposed approach achieved a remarkable level of accuracy, with 91% and 88% on synthetic and real-world English-Bengali code-mixed product review datasets, respectively, outperforming baseline models and proving to be beneficial for future work related to code-mixed sentiment analysis tasks.
Keywords:
BiLSTM, BiGRU, sentiment analysis, ensemble learningReferences
K. D. S. Devi, V. Sireesha, C. Sudha, M. Ravisankar, and P. D. K. Reddy, "A Novel Approach to Sentiment Analysis using GMM-Enhanced N-gram LSTM Networks," Engineering, Technology & Applied Science Research, vol. 15, no. 3, pp. 23068–23073, June 2025.
K. Natarajan, V. Nirmalrani, S. Gowri, R. G. Franklin, D. Poornima, and J. Jabez, "Leveraging Word2Vec-Enhanced CNN-LSTM Hybrid Architecture for Sentiment Analysis in E-Commerce Product Reviews," International Journal of Electrical and Computer Engineering Systems, vol. 17, no. 1, pp. 1–10, Jan. 2026.
L. G. Atlas et al., "A modernized approach to sentiment analysis of product reviews using BiGRU and RNN based LSTM deep learning models," Scientific Reports, vol. 15, no. 1, May 2025, Art. no. 16642.
R. S. Jagdale, V. S. Shirsat, and S. N. Deshmukh, "Sentiment Analysis on Product Reviews Using Machine Learning Techniques," in Cognitive Informatics and Soft Computing, 2019, pp. 639–647.
R. Mahajan, A. S. More, and U. Shah, "Navigating Emotion in Code-Mixed Languages: Performance of ML and DL Models on Hindi-English Text," Procedia Computer Science, vol. 258, pp. 4029–4037, Jan. 2025.
R. Srinivasan and C. N. Subalalitha, "Sentimental analysis from imbalanced code-mixed data using machine learning approaches," Distributed and Parallel Databases, vol. 41, no. 1, pp. 37–52, June 2023.
S. S. Bhagya, G. L. Hrishiraj, and S. NaderaBeevi, "Code-Mixed Sentiment Corpus of Customer Reviews in Malayalam and English for Enhanced Local Market Analysis in India," in 2024 International Conference on Emerging Smart Computing and Informatics (ESCI), Pune, India, Mar. 2024, pp. 1–7.
M. Tareq, Md. F. Islam, S. Deb, S. Rahman, and A. A. Mahmud, "Data-Augmentation for Bangla-English Code-Mixed Sentiment Analysis: Enhancing Cross Linguistic Contextual Understanding," IEEE Access, vol. 11, pp. 51657–51671, 2023.
B. Sultana and K. A. Mamun, "Enhancing Bangla-English Code-Mixed Sentiment Analysis with Cross-Lingual Word Replacement and Data Augmentation," in 2024 6th International Conference on Electrical Engineering and Information & Communication Technology (ICEEICT), Dhaka, Bangladesh, Feb. 2024, pp. 652–657.
M. A. Nezami and M. N. Huda, "Public sentiment identification in social media for code-mixed English-Bangla and Banglish text analysis with machine learning," in Sixth International Conference on Image, Video Processing, and Artificial Intelligence (IVPAI 2024), Kuala Lumpur, Malaysia, Sept. 2024, vol. 13225, pp. 114–119.
F. Hossain, N. Jahan, and J. Abadin, "Sentiment Analysis of Bangla-English Code-Mixed and Transliterated Social Media Comments Using Machine Learning," Journal of Tianjin University Science and Technology, Nov. 2023.
D. Barua and T. S. Walia, "A Deep Learning–Based Framework for Dataset Creation and Sentiment Classification of English–Bengali Code-Mixed Texts," Engineering, Technology & Applied Science Research, vol. 16, no. 1, pp. 31653–31661, Feb. 2026.
"DaliaBarua/En-Bn-Code-Mixed-Two-Class-Sentiment-Dataset." Hugging Face, [Online]. Available: https://huggingface.co/datasets/DaliaBarua/En-Bn-Code-Mixed-Two-Class-Sentiment-Dataset.
M. R. A. Rashid, K. F. Hasan, R. Hasan, A. Das, M. Sultana, and M. Hasan, "A comprehensive dataset for sentiment and emotion classification from Bangladesh e-commerce reviews," Data in Brief, vol. 53, Apr. 2024, Art. no. 110052.
M. R. A. Rashid, K. F. Hasan, M. R. Hasan, A. Das, and M. Sultana, "A Multilabel Multiclass Sentiment and Emotion Dataset from Bangladeshi E-Commerce Reviews," vol. 1, Nov. 2023.
Gaurav and S. K. Mogha, "Advanced Deep Learning Models for Analyzing Sentiments in Code-Mixed Hinglish Video Comments," in Proceedings of the International Conference on AI and Robotics, Astana, Kazakhstan, 2026, pp. 223–235
Downloads
How to Cite
License
Copyright (c) 2026 Dalia Barua, Tarandeep Singh Walia
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain the copyright and grant the journal the right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) after its publication in ETASR with an acknowledgement of its initial publication in this journal.
