Constructing Boosting algorithms from SVMs: an application to one-class classification.

Empirical Inference

Bernhard Schölkopf

Director

Empirical Inference

Gunnar Rätsch

We show via an equivalence of mathematical programs that a support vector (SV) algorithm can be translated into an equivalent boosting-like algorithm and vice versa. We exemplify this translation procedure for a new algorithmone-class leveragingstarting from the one-class support vector machine (1-SVM). This is a first step toward unsupervised learning in a boosting framework. Building on so-called barrier methods known from the theory of constrained optimization, it returns a function, written as a convex combination of base hypotheses, that characterizes whether a given test point is likely to have been generated from the distribution underlying the training data. Simulations on one-class classification problems demonstrate the usefulness of our approach.

Author(s):	Rätsch, G. and Mika, S. and Schölkopf, B. and Müller, K-R.
Journal:	IEEE Transactions on Pattern Analysis and Machine Intelligence
Volume:	24
Number (issue):	9
Pages:	1184-1199
Year:	2002
Month:	September
Day:	0

BibTeX Type:	Article (article)

DOI:	10.1109/TPAMI.2002.1033211

Digital:	0
Electronic Archiving:	grant_archive
Language:	en
Organization:	Max-Planck-Gesellschaft
School:	Biologische Kybernetik

BibTeX

@article{972,
  title = {Constructing Boosting algorithms from SVMs: an application to one-class classification.},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  abstract = {We show via an equivalence of mathematical programs that a support vector (SV) algorithm can be translated into an equivalent boosting-like algorithm and vice versa. We exemplify this translation procedure for a new algorithmone-class leveragingstarting from the one-class support vector machine (1-SVM). This is a first step toward unsupervised learning in a boosting framework. Building on so-called barrier methods known from the theory of constrained optimization, it returns a function, written as a convex combination of base hypotheses, that characterizes whether a given test point is likely to have been generated from the distribution underlying the training data. Simulations on one-class classification problems demonstrate the usefulness of our approach.},
  volume = {24},
  number = {9},
  pages = {1184-1199},
  organization = {Max-Planck-Gesellschaft},
  school = {Biologische Kybernetik},
  month = sep,
  year = {2002},
  author = {R{\"a}tsch, G. and Mika, S. and Sch{\"o}lkopf, B. and M{\"u}ller, K-R.},
  doi = {10.1109/TPAMI.2002.1033211},
  month_numeric = {9}
}