Support vector machine regression algorithm based on chunking incremental learning

Authors:
Jiang Jingqing;Song Chuyi;Wu Chunguo;Marchese Maurizio;Liang Yangchun
Affiliations:
College of Computer Science and Technology, Jilin University, Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education, Changchun, China;College of Mathematics and Computer Science, Inner Mongolia University for Nationalities, Tongliao, China;College of Computer Science and Technology, Jilin University, Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education, Changchun, China;Department of Information and Communication Technology, University of Trento, Povo (TN), Italy;College of Computer Science and Technology, Jilin University, Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education, Changchun, China
Venue:
ICCS'06 Proceedings of the 6th international conference on Computational Science - Volume Part I
Year:
2006

Citing 4
Cited 1

Making large-scale support vector machine learning practical

Advances in kernel methods
Fast training of support vector machines using sequential minimal optimization

Advances in kernel methods
Least Squares Support Vector Machine Classifiers

Neural Processing Letters
Efficient SVM Regression Training with SMO

Machine Learning

A heuristic weight-setting strategy and iteratively updating algorithm for weighted least-squares support vector regression

Neurocomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

On the basis of least squares support vector machine regression (LSSVR), an adaptive and iterative support vector machine regression algorithm based on chunking incremental learning (CISVR) is presented in this paper. CISVR is an iterative algorithm and the samples are added to the working set in batches. The inverse of the matrix of coefficients from previous iteration is used to calculate the regression parameters. Therefore, the proposed approach permits to avoid the calculation of the inverse of a large-scale matrix and improves the learning speed of the algorithm. Support vectors are selected adaptively in the iteration to maintain the sparseness. Experimental results show that the learning speed of CISVR is improved greatly compared with LSSVR for the similar training accuracy. At the same time the number of the support vectors obtained by the presented algorithm is less than that obtained by LSSVR greatly.