Loading...
Constructing a Robust Least Squares Support Vector Machine Based on Lp-norm and L∞-norm
Jahanmard Hosseinabadi, Maryam | 2022
528
Viewed
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 55160 (02)
- University: Sharif University of Technolog
- Department: Mathematical Sciences
- Advisor(s): Mahdavi Amiri, Nezamoddin
- Abstract:
- In the last decades, Support Vector Machine (SVM) has been used for supervised classification due to its wide applications. In addition, SVM is a mathematical programming tool, that as other optimization-based approaches, turns to be useful for a successful development of supervised classification. Based on current research in the literature, we explain the extension of a developed method for SVM using L2-norm to the more general case of Lp-norms with p > 1. The Kernel function is used frequently in the L2-SVM model, but the multidimensional Kernel is used as a general function in Lp-SVM. Finally, we use Lp-LSSVM model, with 0
- Keywords:
- Support Vector Machine (SVM) ; Mathematical Optimization ; Supervised Learning ; Least Squares Method ; Sequential Minimal Optimization (SMO)