An Improved DAG-SVM for Multi-class Classification

Author

Chen, Peng ; Liu, Shuang

Author_Institution

Dept. of Comput. Sci. & Technol., Neusoft Inst. of Inf., Dalian, China

Volume

1

fYear

2009

fDate

14-16 Aug. 2009

Firstpage

460

Lastpage

462

Abstract

Directed Acyclic Graph-Support Vector Machine (DAG-SVM) is a novel algorithm for multi-class classification. For an N-class problem, it constructs N(N-1)/2 classifiers, one for each pair of classes. Based on SVM decision function, an efficient data structure is used to express the decision node in the graph and an improved decision algorithm is used to find the class of each test sample. This new approach remedies some weakness of the DDAG caused by its structure and its sequence of nodes, and makes the decision faster and more accurate. Experimental results on benchmark dataset show the efficiency and improvement of our method.

Keywords

benchmark testing; classification; directed graphs; support vector machines; SVM decision function; benchmark dataset; decision node; directed acyclic graph support vector machine; multi class classification; Benchmark testing; Classification algorithms; Computer science; Data structures; Educational institutions; Kernel; Machine learning; Statistical learning; Support vector machine classification; Support vector machines; decision node; directed acyclic graph; multi-class classification; support vector machine;

fLanguage

English

Publisher

ieee

Conference_Titel

Natural Computation, 2009. ICNC '09. Fifth International Conference on

Conference_Location

Tianjin

Print_ISBN

978-0-7695-3736-8

Type

conf

DOI

10.1109/ICNC.2009.275

Filename

5366976