publication . Other literature type . Conference object . 2005

A support vector method for multivariate performance measures

Thorsten Joachims;
Open Access
  • Published: 01 Jan 2005
  • Publisher: Association for Computing Machinery (ACM)
Abstract
This paper presents a Support Vector Method for optimizing multivariate nonlinear performance measures like the F 1 -score. Taking a multivariate prediction approach, we give an algorithm with which such multivariate SVMs can be trained in polynomial time for large classes of potentially non-linear performance measures, in particular ROCArea and all measures that can be computed from the contingency table. The conventional classification SVM arises as a special case of our method.
Subjects
free text keywords: Support vector method, Mathematics, Multivariate prediction, Special case, Contingency table, Nonlinear system, Multivariate statistics, Support vector machine, Machine learning, computer.software_genre, computer, Pattern recognition, Time complexity, Artificial intelligence, business.industry, business
Related Organizations
Any information missing or wrong?Report an Issue