On the annealed VC entropy for margin classifiers: A statistical mechanics study

Manfred Opper

On the annealed VC entropy for margin classifiers: A statistical mechanics study

Manfred Opper

Aston University

Research output: Chapter in Book/Published conference output › Chapter

Abstract

Using techniques from Statistical Physics, the annealed VC entropy for hyperplanes in high dimensional spaces is calculated as a function of the margin for a spherical Gaussian distribution of inputs.

Original language	English
Title of host publication	Advances in Kernel Methods - Support vector learning
Editors	Bernhard Scholkopf, Christopher J. C. Burges, Alexander J. Smola
Place of Publication	Cambridge, MA
Publisher	MIT
Pages	117-126
Number of pages	10
ISBN (Print)	0262194163
Publication status	Published - 18 Dec 1998

Bibliographical note

Copyright of the Massachusetts Institute of Technology Press (MIT Press) Available in Google Books

Keywords

annealed VC entropy
hyperplanes
high dimensional spaces
spherical Gaussian distribution

Access to Document

NCRG_99_028.pdf
Copyright of the Massachusetts Institute of Technology Press (MIT Press) Available in Google Books
Accepted author manuscript, 248 KB

Cite this

@inbook{0d086b16d65340dd93555136d328e8e5,

title = "On the annealed VC entropy for margin classifiers: A statistical mechanics study",

abstract = "Using techniques from Statistical Physics, the annealed VC entropy for hyperplanes in high dimensional spaces is calculated as a function of the margin for a spherical Gaussian distribution of inputs.",

keywords = "annealed VC entropy, hyperplanes, high dimensional spaces, spherical Gaussian distribution",

author = "Manfred Opper",

note = "Copyright of the Massachusetts Institute of Technology Press (MIT Press) Available in Google Books",

year = "1998",

month = dec,

day = "18",

language = "English",

isbn = "0262194163",

pages = "117--126",

editor = "Bernhard Scholkopf and Burges, {Christopher J. C.} and Smola, {Alexander J.}",

booktitle = "Advances in Kernel Methods - Support vector learning",

publisher = "MIT",

}

TY - CHAP

T1 - On the annealed VC entropy for margin classifiers: A statistical mechanics study

AU - Opper, Manfred

N1 - Copyright of the Massachusetts Institute of Technology Press (MIT Press) Available in Google Books

PY - 1998/12/18

Y1 - 1998/12/18

N2 - Using techniques from Statistical Physics, the annealed VC entropy for hyperplanes in high dimensional spaces is calculated as a function of the margin for a spherical Gaussian distribution of inputs.

AB - Using techniques from Statistical Physics, the annealed VC entropy for hyperplanes in high dimensional spaces is calculated as a function of the margin for a spherical Gaussian distribution of inputs.

KW - annealed VC entropy

KW - hyperplanes

KW - high dimensional spaces

KW - spherical Gaussian distribution

UR - http://mitpress.mit.edu/catalog/item/default.asp?ttype=2&tid=7841

M3 - Chapter

SN - 0262194163

SP - 117

EP - 126

BT - Advances in Kernel Methods - Support vector learning

A2 - Scholkopf, Bernhard

A2 - Burges, Christopher J. C.

A2 - Smola, Alexander J.

PB - MIT

CY - Cambridge, MA

ER -

On the annealed VC entropy for margin classifiers: A statistical mechanics study

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this