An Efficient Model for Opinion Analysis in Dynamic Data Streams

C. Ubani; E.O. Bennett; V.I.E. Anireh; D. Matthias

doi:https://doi.org/10.47001/JAIET/2026.307003

Authors

C. Ubani

Department of Computer Science, Rivers State University, Port Harcourt, Nigeria

E.O. Bennett

Department of Computer Science, Rivers State University, Port Harcourt, Nigeria

V.I.E. Anireh

Department of Computer Science, Rivers State University, Port Harcourt, Nigeria

D. Matthias

Department of Computer Science, Rivers State University, Port Harcourt, Nigeria

Abstract

This study proposes an adaptive opinion analysis framework for real-time sentiment classification in time-evolving textual data streams. Existing approaches to opinion analysis in dynamic data streams face notable limitations, particularly in managing concept drift, handling imbalanced sentiment distributions, and adapting to evolving feature spaces over time thereby falling short of addressing the full range of challenges present in dynamic streaming environments. These limitations motivated the development of a more robust and efficient model capable of meeting the demands of real-world streaming environments. The methodology integrates a transformer-based contextual embedding model (DistilBERT) with an adaptive autoencoder for feature compression and Synthetic Minority Over-sampling Technique (SMOTE) to address class imbalance in streaming sentiment datasets. Concept drift in dynamic opinion data is monitored using Feature Selection Stability and Drift (FSSD) combined with the ADWIN drift detection algorithm, enabling the system to adapt to evolving sentiment patterns over time. The research adopted object-oriented analysis and design as the system development methodology due to its reusability. Experimental evaluation was conducted on a continuously streamed customer review dataset using an 80:20 training-testing split. The transformer model was trained using the AdamW optimiser with a learning rate of 2×10⁻⁵ over ten epochs, achieving a validation accuracy of 94.50%, precision of 0.93, recall of 0.94, F1-score of 0.94, and ROC-AUC score of 0.79 after drift adaptation. The results demonstrate that the proposed system significantly improves adaptability and classification robustness compared to static feature-based streaming models, particularly in handling feature evolution and imbalanced sentiment distributions in real-time opinion analysis environments.

Keywords

Adaptive Learning Concept Drift Dynamic Data streams DistilBERT Feature Evolution Opinion Analysis SMOTE.

Citation of this Article

C. Ubani, E.O. Bennett, V.I.E. Anireh, & D. Matthias. (2026). An Efficient Model for Opinion Analysis in Dynamic Data Streams. Journal of Artificial Intelligence and Emerging Technologies (JAIET). 3(7), 24-32. Article DOI: https://doi.org/10.47001/JAIET/2026.307003

References

Cortis & Davis, B. (2021). Over a decade of social opinion mining: a systematic review. Artificial Intelligence Review, 54(7), 4873-4965. https://doi.org/10.1007/s10462-021-10030-2

Zhang, L., Wang, S., & Liu, B. (2018). Deep learning for sentiment analysis: a survey. Wiley Interdisciplinary Reviews Data Mining and Knowledge Discovery, 8(4). https://doi.org/10.1002/widm.1253

Bechini, A., Bondielli, A., Ducange, P., Marcelloni, F., &Renda, A. (2021). Addressing event-driven concept drift in twitter stream: a stance detection application. IEEE Access, 9, 77758-77770. https://doi.org/10.1109/access.2021.3083578

Wu, Y., Sharma, K., Seah, C., & Zhang, S. (2023). Sentistream: a co-training framework for adaptive online sentiment analysis in evolving data streams. https://doi.org/10.18653/v1/2023.emnlp-main.380

Susi, E. &Shanthi, A. (2023). Sentiment drift detection and analysis in real time twitter data streams. Computer Systems Science and Engineering, 45(3), 3231-3246. https://doi.org/10.32604/csse.2023.032104

Chung, S. Å, D. (2018). Integrated real-time big data stream sentiment analysis service. Journal of Data Analysis and Information Processing, 06(02), 46-66. https://doi.org/10.4236/jdaip.2018.62004

Chakravarthy, A. (2018). Hybrid architecture for sentiment analysis using deep learning. International Journal of Advanced Research in Computer Science, 9(1), 735-738. https://doi.org/10.26483/ijarcs.v9i1.5388

Redjeki, S. &Widyarto, S. (2022). Comparison of seven machine learning algorithms in the classification of public opinion. Tech-E, 5(2), 143-149. https://doi.org/10.31253/te.v5i1.1046

Khandekar, V. &Shrinath, P. (2022). Hybrid dynamic chunk ensemble model for multi-class data streams. Indonesian Journal of Electrical Engineering and Computer Science, 25(2), 1115. https://doi.org/10.11591/ijeecs.v25.i2.pp1115-1122

Museba, T., Nelwamondo, F., &Ouahada, K. (2021). Ades: a new ensemble diversity-based approach for handling concept drift. Mobile Information Systems, 2021, 1-17. https://doi.org/10.1155/2021/5549300

Krawczyk, B., Minku, L., Gama, J., Stefanowski, J., &Woniak, M. (2017). Ensemble learning for data stream analysis: a survey. Information Fusion, 37, 132-156. https://doi.org/10.1016/j.inffus.2017.02.004

Muslim, A., Benny, A., Refianti, R., Maisyarah, C., & Setiawan, G. (2020). Comparison of accuracy between long short-term memory-deep learning and multinomial logistic regression-machine learning in sentiment analysis on twitter. International Journal of Advanced Computer Science and Applications, 11(2). https://doi.org/10.14569/ijacsa.2020.0110294

Jafseer,K.T., Shailesh, S., & Sreekumar, A. (2023). A feature evolution aware classification framework for streaming data using dynamic autoencoder and ensemble learning. https://doi.org/10.21203/rs.3.rs-2479805/v1

Tuama, M. (2023). Beyond polarity: the potential applications and impacts of sentiment analysis and emotion detection. AKJS, 1(2), 44-51. https://doi.org/10.61710/akjs.v1i2.51

Karanikola, A. (2023). Financial sentiment analysis: Classic methods vs. deep learning models. Intelligent Decision Technologies, 17(4), 893-915. https://doi.org/10.3233/idt-230478