Open source software for training large-scale linear and nonlinear Support Vector Machines within scalar environments. It implements a standard decomposition approach for the dual quadratic programming (QP) formulation, where the solver for the inner QP problems is based on iterative gradient-projection methods. An effective caching strategy is also used to save computational time in kernel evaluations. The main feature is that it can efficiently handle medium-to-large inner QP problems, whilst all the other competing packages can not.
GPDT - Gradient-Projection-based Decomposition Technique
ZANGHIRATI, Gaetano;
2004
Abstract
Open source software for training large-scale linear and nonlinear Support Vector Machines within scalar environments. It implements a standard decomposition approach for the dual quadratic programming (QP) formulation, where the solver for the inner QP problems is based on iterative gradient-projection methods. An effective caching strategy is also used to save computational time in kernel evaluations. The main feature is that it can efficiently handle medium-to-large inner QP problems, whilst all the other competing packages can not.File in questo prodotto:
Non ci sono file associati a questo prodotto.
I documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
