Bounds for Sparse Solutions of K-SVCR Multi-class Classification Model

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F44555601%3A13440%2F23%3A43897570" target="_blank" >RIV/44555601:13440/23:43897570 - isvavai.cz</a>

Result on the web

<a href="https://link.springer.com/chapter/10.1007/978-3-031-24866-5_11#Abs1" target="_blank" >https://link.springer.com/chapter/10.1007/978-3-031-24866-5_11#Abs1</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-031-24866-5_11" target="_blank" >10.1007/978-3-031-24866-5_11</a>

Result language

angličtina

Original language name

Bounds for Sparse Solutions of K-SVCR Multi-class Classification Model

Original language description

The support vector classification-regression machine for k-class classification (K-SVCR) is a novel multi-class classification approach based on the ?1-versus-1-versus-rest? structure. In this work, we suggested an efficient model by proposing the p-norm (0&lt;p&lt;1) instead of the 2-norm for the regularization term in the objective function of K-SVCR that can be used for feature selection. We derived lower bounds for the absolute value of nonzero entries in every local optimal solution of the p-norm based model. Also, we provided upper bounds for the number of nonzero components of the optimal solutions. We explored the link between solution sparsity, regularization parameters, and the p-choice.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Publication year

2023

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

LION 2022: Learning and Intelligent Optimization

ISBN

978-3-031-24865-8

ISSN

—

e-ISSN

—

Number of pages

9

Pages from-to

136-144

Publisher name

Springer, Cham

Place of publication

Switzerland AG

Event location

Milos Island

Event date

Jun 5, 2022

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—