This title appears in the Scientific Report :
2006
Please use the identifier:
http://dx.doi.org/10.1016/j.cam.2005.09.009 in citations.
Efficient Optimization of Support Vector Machine Learning Parameters for Unbalanced Data Sets
Efficient Optimization of Support Vector Machine Learning Parameters for Unbalanced Data Sets
Support vector machines are powerful kernel methods for classification and regression tasks. If trained optimally, they produce excellent separating hyperplanes. The quality of the training, however, depends not only on the given training data but also on additional learning parameters, which are di...
Saved in:
Personal Name(s): | Eitrich, T. |
---|---|
Lang, B. | |
Contributing Institute: |
Zentralinstitut für Angewandte Mathematik; ZAM |
Published in: | Journal of Computational and Applied Mathematics, 196 (2006) S. 425 - 436 |
Imprint: |
Amsterdam [u.a.]
North-Holland
2006
|
Physical Description: |
425 - 436 |
DOI: |
10.1016/j.cam.2005.09.009 |
Document Type: |
Journal Article |
Research Program: |
Scientific Computing |
Series Title: |
Journal of Computational and Applied Mathematics
196 |
Subject (ZB): | |
Publikationsportal JuSER |
Support vector machines are powerful kernel methods for classification and regression tasks. If trained optimally, they produce excellent separating hyperplanes. The quality of the training, however, depends not only on the given training data but also on additional learning parameters, which are difficult to adjust, in particular for unbalanced datasets. Traditionally, grid search techniques have been used for determining suitable values for these parameters. In this paper, we propose an automated approach to adjusting the learning parameters using a derivative-free numerical optimizer. To make the optimization process more efficient, a new sensitive quality measure is introduced. Numerical tests with a well-known dataset show that our approach can produce support vector machines that are very well tuned to their classification tasks. (c) 2005 Elsevier B.V. All rights reserved. |