International Journal of Innovative Research in Computer and Communication Engineering
ISSN Approved Journal | Impact factor: 8.771 | ESTD: 2013 | Follows UGC CARE Journal Norms and Guidelines
| Monthly, Peer-Reviewed, Refereed, Scholarly, Multidisciplinary and Open Access Journal | High Impact Factor 8.771 (Calculated by Google Scholar and Semantic Scholar | AI-Powered Research Tool | Indexing in all Major Database & Metadata, Citation Generator | Digital Object Identifier (DOI) |
| TITLE | Self-Supervised Learning Models for Scalable Data Stream Mining |
|---|---|
| ABSTRACT | The emergence of an exponentially increasing amount of real-time data streams originating from IoT devices, transactions, and industry sensors has proven too much to manage using standard supervised learning methods requiring large datasets with manual labeling. The application of Self-Supervised Learning (SSL) can help overcome these limitations through the extraction of useful features from unlabeled sequential data. This paper provides a detailed overview of SSL methods utilized in data stream mining and includes a review of the latest research developments in contrastive learning, generative replay techniques, and semi-supervised frameworks. Three exemplary self-supervised learning methods (Cross-Domain Predictive and Contextual Contrasting (CDPCC), Graph-Theory-based Semi-supervised Self-training (GTSS), and Hierarchical Contrastive Learning with Importance-aware Resolution Selection (IARS)) were tested on a range of problems including fault detection, streaming classification, and time series. Our study shows that CDPCC provides fault detection at 95.6% accuracy while only using 50% of labeled data, GTSS classifies data with 99.6% accuracy when 10% is labeled, and IARS cuts down on training time by 40% without damaging the integrity of the model. |
| AUTHOR | G. MOHANA PRIYA, K. PREETHA Department of Information Technology, Coimbatore Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India |
| VOLUME | 184 |
| DOI | DOI: 10.15680/IJIRCCE.2026.1405059 |
| pdf/59_Self-Supervised Learning Models for Scalable Data Stream Mining.pdf | |
| KEYWORDS | |
| References | [1]. S. Zhu, "Continual Self-Supervised Learning with Diffusion Models," NAISS Small Compute Project, Uppsala University, Sweden, 2025-2026. [2]. L. Zólyomi, O. Smirnov, et al., "Towards Unified Approaches in Self-Supervised Event Stream Modeling: Progress and Prospects," arXiv preprint arXiv:2502.04899, 2025. [3]. Y. Liu, K. Wang, and H. Zhang, "Time series representation learning via cross-domain predictive and contextual contrasting: Application to fault detection," Engineering Applications of Artificial Intelligence, vol. 138, 2025. [4]. N. Samadi, J. Tanha, and M. Jalili, "Graph theory-based semi-supervised self-training for data stream classification and emerging class detection," Information Sciences, vol. 698, p. 121762, 2025. [5]. H.-L. Nguyen, W.-K. Ng, and Y.-K. Woon, "Concurrent Semi-Supervised Learning with Active Learning of Data Streams," in Proc. TL-DKS, vol. 8, 2013. [6]. J. A. Perez, "Efficient Hierarchical Contrastive Self-supervising Learning for Time Series Classification via Importance-aware Resolution Selection," in Proc. IEEE BigData 2024, Washington, DC, pp. 880-889, 2024. [7]. N. Samadi, J. Tanha, and M. Jalili, "GTSS: Graph-Theory-based Semi-supervised Self-training for Data Streams," Information Sciences, 2024 (available online). [8]. G. Cignoni, S. Magistri, A. Carta, and A. D. Bagdanov, "SOLAR: Adaptive Online Continual Self-Supervised Learning," arXiv preprint, 2026. [9]. "Garuda - Garba Rujukan Digital," Indonesian Publication Index, 2025 (repository metadata). [10]. K. Rao, "Selective self-training semi-supervised classification for data streams," Proceeding of the Electrical Engineering, Computer Science, and Informatics, 2024. |
