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

by: G Ratsch, S Mika, B Scholkopf, KR Muller

Pattern Analysis and Machine Intelligence, IEEE Transactions on, Vol. 24, No. 9. (September 2002), pp. 1184-1199.

View FullText article

Reviews [Write a review of this article]

There are no reviews of this article

Find related articles from these CiteULike users

AbnerCYH

Find related articles with these CiteULike tags

algorithms, kdd, optimization

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 algorithm—one-class leveraging—starting 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.

@article{citeulike:2782486, 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 algorithm—one-class leveraging—starting 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 = {Ratsch, G. and Mika, S. and Scholkopf, B. and Muller, K. R. }, citeulike-article-id = {2782486}, doi = {10.1109/TPAMI.2002.1033211}, journal = {Pattern Analysis and Machine Intelligence, IEEE Transactions on}, keywords = {algorithms, kdd, optimization}, month = {September}, number = {9}, pages = {1184--1199}, posted-at = {2008-05-10 08:08:31}, priority = {5}, title = {Constructing boosting algorithms from SVMs: an application to one-class classification}, url = {http://dx.doi.org/10.1109/TPAMI.2002.1033211}, volume = {24}, year = {2002} }

RIS record

TY - JOUR ID - citeulike:2782486 L3 - citeulike-article-id:2782486 TI - Constructing boosting algorithms from SVMs: an application to one-class classification JF - Pattern Analysis and Machine Intelligence, IEEE Transactions on VL - 24 IS - 9 SP - 1184 EP - 1199 N2 - 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 algorithm—one-class leveraging—starting 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. KW - algorithms KW - kdd KW - optimization AU - Ratsch, G AU - Mika, S AU - Scholkopf, B AU - Muller, KR PY - 2002/09// UR - http://dx.doi.org/10.1109/TPAMI.2002.1033211 ER -

RIS BibTeX