An SOM-Based Decoding Algorithm for Multi-class SVM

Author

Tao, Xiaoyan ; Ji, Hongbing

Author_Institution

Sch. of Electron. Eng., Xidian Univ., Xi´´an

Volume

1

fYear

2005

Firstpage

270

Lastpage

273

Abstract

How to process multi-class problem with SVM is one of the present research focuses. In general, the classification process for the multi-class SVM includes two parts: the encoding and decoding strategy. Specially, we address the decoding problem which concerns how to map the outputs into class codewords. In this paper an SOM-based decoding algorithm for multi-class SVM is presented, which directly use the output magnitude to classify the data in combination with the nonlinearity and topological ordering of the SOM. The experiments on the Yale and ORL face databases show the advantage of the new method over the widely-used Hamming decoding scheme

Keywords

decoding; learning (artificial intelligence); self-organising feature maps; support vector machines; visual databases; Hamming decoding scheme; ORL face databases; decoding algorithm; multi-class problem; self-organizing maps; support vector machine; Databases; Decoding; Electrical capacitance tomography; Encoding; Error correction codes; Machine learning; Machine learning algorithms; Optimization methods; Support vector machine classification; Support vector machines;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks and Brain, 2005. ICNN&B '05. International Conference on

Conference_Location

Beijing

Print_ISBN

0-7803-9422-4

Type

conf

DOI

10.1109/ICNNB.2005.1614613

Filename

1614613