A New Algorithm for SVM Incremental Learning

Author

Wang, Xiaodan ; Zheng, Chunying ; Wu, Chongming ; Wang, Wei

Author_Institution

Dept. of Comput. Eng., Air Force Eng. Univ.

Volume

3

fYear

2006

fDate

16-20 Nov. 2006

Abstract

Based on analyzing the relationship between the Karush-Kuhn-Tucker (KKT) conditions of support vector machine and the distribution of the training samples, the possible changes of support vector set after new samples are added to training set was analyzed, and the generalized Karush-Kuhn-Tucker conditions was defined. Based on the classification equivalence between the previous training set and the newly added training set, a new algorithm for SVM incremental learning is proposed. With the presented algorithm, the useless sample is discarded and useful information in training samples is accumulated. Experimental results with the standard datasets indicate the effectiveness of the proposed algorithm

Keywords

learning (artificial intelligence); signal classification; support vector machines; Karush-Kuhn-Tucker conditions; SVM incremental learning; support vector machine; training set; Algorithm design and analysis; Distributed computing; Kernel; Lagrangian functions; Machine learning; Machine learning algorithms; Military computing; Pattern recognition; Support vector machine classification; Support vector machines;

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing, 2006 8th International Conference on

Conference_Location

Beijing

Print_ISBN

0-7803-9736-3

Electronic_ISBN

0-7803-9736-3

Type

conf

DOI

10.1109/ICOSP.2006.345821

Filename

4129190