On the VC Dimension of Bounded Margin Classifiers

In this paper we prove a result that is fundamental to the generalization properties of Vapnik's support vector machines and other large margin classifiers. In particular, we prove that the minimum margin over all dichotomies of k ≤ n + 1 points inside a unit ball in R^sup n^ is maximized when...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Machine learning 2001-10, Vol.45 (1), p.33-44
Hauptverfasser: Hush, Don, Scovel, Clint
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:In this paper we prove a result that is fundamental to the generalization properties of Vapnik's support vector machines and other large margin classifiers. In particular, we prove that the minimum margin over all dichotomies of k ≤ n + 1 points inside a unit ball in R^sup n^ is maximized when the points form a regular simplex on the unit sphere. We also provide an alternative proof directly in the framework of level fat shattering.[PUBLICATION ABSTRACT]
ISSN:0885-6125
1573-0565
DOI:10.1023/A:1010971905232