Exploring Proximal and Level Bundle Methods for Regularized Support Vector Machines

Rusnák, Radovan (2023) Exploring Proximal and Level Bundle Methods for Regularized Support Vector Machines. Bachelor's Thesis, Mathematics.

Preview	Text bMATH_2023_RadovanRusnák.pdf Download (858kB) | Preview
	Text toestemming.pdf Restricted to Registered users only Download (130kB)

Abstract

In this paper, we study the Proximal and Level Bundle Methods and their effectiveness in handling the convex non-smooth optimization problem associated with Regularized Support Vector Machine. Many of the traditional optimization algorithms fail to handle non-smooth objective functions, resulting in unstable behaviour and slow convergence. Furthermore, in practical computational problems, dealing with large datasets requires efficient memory management during computations. Bundle Methods tackle these challenges through a distinctive compression mechanism that enables working with a bounded amount of information at each iteration and guaranteed stability. The versatility of Bundle Methods makes them highly applicable to a wide range of optimization problems, and they are currently one of the most promising and popular optimization techniques in many practical applications.

Item Type:	Thesis (Bachelor's Thesis)
Supervisor name:	Peypouquet, J.G. and Waarde, H.J. van
Degree programme:	Mathematics
Thesis type:	Bachelor's Thesis
Language:	English
Date Deposited:	10 Jul 2023 09:38
Last Modified:	10 Jul 2023 09:38
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/30403

Actions (login required)

View Item