Improvements to Platt's SMO Algorithm for SVM Classifier Design

Improvements to Platt's SMO Algorithm for SVM Classifier Design (1999)

Cached

Download Links

[guppy.mpe.nus.edu.sg]

Other Repositories/Bibliography

DBLP

by S.S. Keerthi , S. K. Shevade , C. Bhattacharyya , K. R. K. Murthy

Citations:

139 - 10 self

Versions

Version History

Metadata Version 1

Datum	Value	Source
TITLE	Imrovements to Platt's SMO Algorithm for SVM Classi er Design 1	SVM HeaderParse 0.1
AUTHOR NAME	S. S. Keerthi	SVM HeaderParse 0.1
AUTHOR AFFIL	Control Division; Dept. of Mechanical and Production Engineering; National University of Singapore	SVM HeaderParse 0.2
AUTHOR ADDR	Singapore-119260	SVM HeaderParse 0.1
AUTHOR NAME	S. K. Shevade	SVM HeaderParse 0.1
AUTHOR AFFIL	Control Division; Dept. of Mechanical and Production Engineering; National University of Singapore	SVM HeaderParse 0.2
AUTHOR ADDR	Singapore-119260	SVM HeaderParse 0.1
AUTHOR NAME	C. Bhattacharyya	SVM HeaderParse 0.1
AUTHOR AFFIL	Control Division; Dept. of Mechanical and Production Engineering; National University of Singapore	SVM HeaderParse 0.2
AUTHOR ADDR	Singapore-119260	SVM HeaderParse 0.1
AUTHOR NAME	K. R. K. Murthy	SVM HeaderParse 0.1
AUTHOR AFFIL	Control Division; Dept. of Mechanical and Production Engineering; National University of Singapore	SVM HeaderParse 0.2
AUTHOR ADDR	Singapore-119260	SVM HeaderParse 0.1
ABSTRACT	A revised version of this report is under preparation for submission to a journal. We welcome any comments and suggestions for improving this report. This paper points out an important source of confusion and ineciency in Platt's Sequential Minimal Optimization (SMO) algorithm that is caused by the use of a single threshold value. Using clues from the KKT conditions for the dual problem, two threshold parameters are employed to derive modications of SMO. These modied algorithms perform signicantly faster than the original SMO on all benchmark datasets tried. 1	SVM HeaderParse 0.1
CITATIONS	5 found	ParsCit 1.0