A Sparse Twin SVM for multi-classification problems

Journal Title: INTERNATIONAL JOURNAL OF MATHEMATICS TRENDS AND TECHNOLOGY - Year 2013, Vol 4, Issue 3

Abstract

 We propose Sparse TSVM, a multi-class SVM classifier that determines k nonparallel planes by solving k related SVM-type problems. The Sparse TSVM promotes Twin SVM to one-versus-rest approach. And it capture classes' main feature better with the sparse algorithm. On several benchmark data sets, Sparse TSVM is not only fast, but shows good generalization.

Authors and Affiliations

HONG-XING YAO

Keywords

Related Articles

 Magic and Bimagic Labeling for Disconnected Graphs

 An edge magic total labeling of a graph G V E ( , ) with p vertices and q edges is a bijection f from the set of vertices and edges to 1,2,..., p q  such that for every edge uv in E, f u f uv f v ( ) ( ) ( )  ...

 Hematocrit effects of the axisymmetric blood flow through an artery with stenosis arteries

 When the nature of blood flow changes from its usual state to a disturbed flow condition due to the presence of a stenosis in an artery. Then it has been suggested that the deposits of cholesterol on the arterial w...

Boolean Algebras of Pre A*-Algebra

In this paper, we obtain Boolean algebras , from Pre A*-algebra , and proved that , and are isomorphic to each other

 Equivalence Of Super Magic Labelings

 Let G (V, E) be a graph with p vertices and q edges. A bijection f : V E  {1,2,…., p+q} is called a super magic labeling of a graph G, if f(V) = {1, 2,…., p} and for any edge xyE, f(x)+f(y)+f(xy) = c(f), a const...

Mathematical Technique Based Speaker Recognition for Total Security System

Today security system is rapidly growing and updating simultaneously due to advance of technology. This growth in electronic transactions results in a rise of demand for fast and accurate user identification and authenti...

Download PDF file
  • EP ID EP120625
  • DOI -
  • Views 96
  • Downloads 0

How To Cite

HONG-XING YAO (2013).  A Sparse Twin SVM for multi-classification problems. INTERNATIONAL JOURNAL OF MATHEMATICS TRENDS AND TECHNOLOGY, 4(3), 41-52. https://europub.co.uk/articles/-A-120625