Title :
Error-centre-based algorithm for support vector machine training
Author :
Meng, L. ; Wu, Q.H.
Author_Institution :
Dept. of Electr. Eng. & Electron., Liverpool Univ., UK
fDate :
3/28/2002 12:00:00 AM
Abstract :
A new error-centre-based algorithm for support vector machine training is proposed. While the standard techniques require a memory space growing quadratically with the problem size, the new algorithm requires only a linearly growing memory space and may thus be applied to much larger data sets compared with the standard techniques
Keywords :
image segmentation; learning automata; quadratic programming; data sets; error-centre-based algorithm; linearly growing memory space; support vector machine training;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20020240
