Online Semi-Supervised independent support vector machines
International Journal of Approximate Reasoning, cilt.194, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 194
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.ijar.2026.109658
- Dergi Adı: International Journal of Approximate Reasoning
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC, MathSciNet, zbMATH, DIALNET, Academic Search Ultimate (EBSCO)
- Anahtar Kelimeler: Incremental learning, Linear independence, Machine learning, Manifold regularization, Online learning, Semi-supervised learning, Support vector machine
- Hacettepe Üniversitesi Adresli: Evet
Özet
Support Vector Machines (SVMs) have garnered significant interest due to their high performance and broad range of applications. The standard SVM algorithm is a supervised learning system that relies on sufficient labeled data to determine the optimal decision hyperplane. However, acquiring labeled data is often costly or challenging in many real-world applications. Furthermore, the standard SVM approach is inefficient for processing online data streams because it operates as a batch-learning algorithm, necessitating retraining the classifier whenever new data is incorporated. In this paper, we introduce a novel online algorithm, the Semi-Supervised Independent Support Vector Machine (OS2ISVM), which enhances the efficiency of SVMs in online semi-supervised learning contexts. This method effectively manages the number of Support Vectors (SVs) by constructing an independent linear set of training data, thereby significantly reducing the complexity of classification. The utilization of the Newton method facilitates rapid convergence and improves computational speed. Experimental results demonstrate that OS2ISVM effectively reduces computational load while maintaining high classification accuracy. This offers a practical solution for real-world applications.