A hybrid classifier based on boxes and nearest neighbors

Martin Anthony; Joel Ratsaby

doi:10.1016/j.dam.2014.02.018

A hybrid classifier based on boxes and nearest neighbors

Martin Anthony, Joel Ratsaby

Research output: Contribution to journal › Article › peer-review

6 Scopus citations

Abstract

In this paper we analyse the generalization performance of a type of binary classifier defined on the unit cube. This classifier combines some of the aspects of the standard methods that have been used in the logical analysis of data (LAD) and geometric classifiers, with a nearest-neighbor paradigm. We assess the predictive performance of the new classifier in learning from a sample, obtaining generalization error bounds that improve as a measure of 'robustness' of the classifier on the training sample increases.

Original language	English
Pages (from-to)	1-11
Number of pages	11
Journal	Discrete Applied Mathematics
Volume	172
DOIs	https://doi.org/10.1016/j.dam.2014.02.018
State	Published - 31 Jul 2014
Externally published	Yes

Keywords

Generalization error
LAD methods
Large margin learning
Learning algorithms
Logical analysis of data
Machine learning

ASJC Scopus subject areas

Discrete Mathematics and Combinatorics
Applied Mathematics

Access to Document

10.1016/j.dam.2014.02.018

Cite this

@article{350818a5ddbe44a18ba05aa56a0536d9,

title = "A hybrid classifier based on boxes and nearest neighbors",

abstract = "In this paper we analyse the generalization performance of a type of binary classifier defined on the unit cube. This classifier combines some of the aspects of the standard methods that have been used in the logical analysis of data (LAD) and geometric classifiers, with a nearest-neighbor paradigm. We assess the predictive performance of the new classifier in learning from a sample, obtaining generalization error bounds that improve as a measure of 'robustness' of the classifier on the training sample increases.",

keywords = "Generalization error, LAD methods, Large margin learning, Learning algorithms, Logical analysis of data, Machine learning",

author = "Martin Anthony and Joel Ratsaby",

note = "Funding Information: This work was supported in part by the IST Programme of the European Community, under the PASCAL2 Network of Excellence, IST-2007-216886. Joel Ratsaby acknowledges the support of the Department of Computer Science at UCL.",

year = "2014",

month = jul,

day = "31",

doi = "10.1016/j.dam.2014.02.018",

language = "English",

volume = "172",

pages = "1--11",

journal = "Discrete Applied Mathematics",

issn = "0166-218X",

publisher = "Elsevier B.V.",

}

TY - JOUR

T1 - A hybrid classifier based on boxes and nearest neighbors

AU - Anthony, Martin

AU - Ratsaby, Joel

N1 - Funding Information: This work was supported in part by the IST Programme of the European Community, under the PASCAL2 Network of Excellence, IST-2007-216886. Joel Ratsaby acknowledges the support of the Department of Computer Science at UCL.

PY - 2014/7/31

Y1 - 2014/7/31

N2 - In this paper we analyse the generalization performance of a type of binary classifier defined on the unit cube. This classifier combines some of the aspects of the standard methods that have been used in the logical analysis of data (LAD) and geometric classifiers, with a nearest-neighbor paradigm. We assess the predictive performance of the new classifier in learning from a sample, obtaining generalization error bounds that improve as a measure of 'robustness' of the classifier on the training sample increases.

AB - In this paper we analyse the generalization performance of a type of binary classifier defined on the unit cube. This classifier combines some of the aspects of the standard methods that have been used in the logical analysis of data (LAD) and geometric classifiers, with a nearest-neighbor paradigm. We assess the predictive performance of the new classifier in learning from a sample, obtaining generalization error bounds that improve as a measure of 'robustness' of the classifier on the training sample increases.

KW - Generalization error

KW - LAD methods

KW - Large margin learning

KW - Learning algorithms

KW - Logical analysis of data

KW - Machine learning

UR - http://www.scopus.com/inward/record.url?scp=84899639989&partnerID=8YFLogxK

U2 - 10.1016/j.dam.2014.02.018

DO - 10.1016/j.dam.2014.02.018

M3 - Article

AN - SCOPUS:84899639989

SN - 0166-218X

VL - 172

SP - 1

EP - 11

JO - Discrete Applied Mathematics

JF - Discrete Applied Mathematics

ER -

A hybrid classifier based on boxes and nearest neighbors

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this