Article ID: | iaor2007420 |
Country: | United Kingdom |
Volume: | 21 |
Issue: | 5 |
Start Page Number: | 693 |
End Page Number: | 706 |
Publication Date: | Oct 2006 |
Journal: | Optimization Methods & Software |
Authors: | Jeyakumar V., Womersley R.S., Ormerod J. |
Keywords: | programming (semidefinite) |
In this paper, we present knowledge-based support vector machine (SVM) classifiers using semidefinite linear programming. SVMs are an optimization-based solution method for large-scale data classification problems. Knowledge-based SVM classifiers, where prior knowledge is in the form of ellipsoidal constraints, result in a semidefinite linear program with a set containment constraint. These problems are reformulated as standard semidefinite linear programming problems by the application of a dual characterization of the set containment under a mild regularity condition. The reformulated semidefinite linear program is solved by the publicly available solvers. Computational results show that prior knowledge can often improve correctness of the classifier.