Sensitivity of multi-dimensional Bayesian classifiers

Janneke H. Bolt; Silja Renooij

doi:10.3233/978-1-61499-419-0-971

Sensitivity of multi-dimensional Bayesian classifiers

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Abstract

One-dimensional Bayesian network classifiers (OBCs) are popular
tools for classification [2]. An OBC is a Bayesian network [4] consisting
of just a single class variable and several feature variables.
Multi-dimensional Bayesian network classifiers (MBCs) were introduced
to generalise OBCs to multiple class variables [1, 6].
Classification performance of OBCs is known to be rather good.
Experimental results that support this observation were substantiated
by a study of the sensitivity properties of naive OBCs [5]. In this
paper we investigate the sensitivity of MBCs. We present sensitivity
functions for the outcome probabilities of interest of an MBC and use
these functions to study the sensitivity value. This value captures the
sensitivity of an output probability to small changes in a parameter.
We compare MBCs to OBCs in this respect and conclude that an
MBC will on average be even more robust to parameter changes than
an OBC.

Original language	English
Title of host publication	Proceedings of the 21st European Conference on Artificial Intelligence
Subtitle of host publication	(ECAI)
Editors	Torsten Schaub, Gerhard Friedrich, Barry O'Sullivan
Publisher	IOS Press
Pages	971-972
ISBN (Electronic)	978-1-61499-419-0
ISBN (Print)	978-1-61499-418-3
DOIs	https://doi.org/10.3233/978-1-61499-419-0-971
Publication status	Published - 2014
Event	ECAI 2014 - Prague, Czech Republic Duration: 18 Aug 2014 → 22 Aug 2014

Publication series

Name	Frontiers in Artificial Intelligence and Applications
Publisher	IOS Press
Volume	263
ISSN (Print)	0922-6389
ISSN (Electronic)	1879-8314

Conference

Conference	ECAI 2014
Country/Territory	Czech Republic
City	Prague
Period	18/08/14 → 22/08/14

Access to Document

10.3233/978-1-61499-419-0-971

1Final published version, 447 KB

Cite this

@inproceedings{cdb7eac80521448d9ca8417b7723a899,

title = "Sensitivity of multi-dimensional Bayesian classifiers",

abstract = "One-dimensional Bayesian network classifiers (OBCs) are populartools for classification [2]. An OBC is a Bayesian network [4] consistingof just a single class variable and several feature variables.Multi-dimensional Bayesian network classifiers (MBCs) were introducedto generalise OBCs to multiple class variables [1, 6].Classification performance of OBCs is known to be rather good.Experimental results that support this observation were substantiatedby a study of the sensitivity properties of naive OBCs [5]. In thispaper we investigate the sensitivity of MBCs. We present sensitivityfunctions for the outcome probabilities of interest of an MBC and usethese functions to study the sensitivity value. This value captures thesensitivity of an output probability to small changes in a parameter.We compare MBCs to OBCs in this respect and conclude that anMBC will on average be even more robust to parameter changes thanan OBC.",

author = "Bolt, {Janneke H.} and Silja Renooij",

year = "2014",

doi = "10.3233/978-1-61499-419-0-971",

language = "English",

isbn = "978-1-61499-418-3",

series = "Frontiers in Artificial Intelligence and Applications ",

publisher = "IOS Press",

pages = "971--972",

editor = "Torsten Schaub and Gerhard Friedrich and Barry O'Sullivan",

booktitle = "Proceedings of the 21st European Conference on Artificial Intelligence",

address = "Netherlands",

note = "ECAI 2014 ; Conference date: 18-08-2014 Through 22-08-2014",

}

Bolt, JH & Renooij, S 2014, Sensitivity of multi-dimensional Bayesian classifiers. in T Schaub, G Friedrich & B O'Sullivan (eds), Proceedings of the 21st European Conference on Artificial Intelligence: (ECAI). Frontiers in Artificial Intelligence and Applications , vol. 263, IOS Press, pp. 971-972, ECAI 2014, Prague, Czech Republic, 18/08/14. https://doi.org/10.3233/978-1-61499-419-0-971

Sensitivity of multi-dimensional Bayesian classifiers. / Bolt, Janneke H.; Renooij, Silja.
Proceedings of the 21st European Conference on Artificial Intelligence: (ECAI). ed. / Torsten Schaub; Gerhard Friedrich; Barry O'Sullivan. IOS Press, 2014. p. 971-972 (Frontiers in Artificial Intelligence and Applications ; Vol. 263).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Sensitivity of multi-dimensional Bayesian classifiers

AU - Bolt, Janneke H.

AU - Renooij, Silja

PY - 2014

Y1 - 2014

N2 - One-dimensional Bayesian network classifiers (OBCs) are populartools for classification [2]. An OBC is a Bayesian network [4] consistingof just a single class variable and several feature variables.Multi-dimensional Bayesian network classifiers (MBCs) were introducedto generalise OBCs to multiple class variables [1, 6].Classification performance of OBCs is known to be rather good.Experimental results that support this observation were substantiatedby a study of the sensitivity properties of naive OBCs [5]. In thispaper we investigate the sensitivity of MBCs. We present sensitivityfunctions for the outcome probabilities of interest of an MBC and usethese functions to study the sensitivity value. This value captures thesensitivity of an output probability to small changes in a parameter.We compare MBCs to OBCs in this respect and conclude that anMBC will on average be even more robust to parameter changes thanan OBC.

AB - One-dimensional Bayesian network classifiers (OBCs) are populartools for classification [2]. An OBC is a Bayesian network [4] consistingof just a single class variable and several feature variables.Multi-dimensional Bayesian network classifiers (MBCs) were introducedto generalise OBCs to multiple class variables [1, 6].Classification performance of OBCs is known to be rather good.Experimental results that support this observation were substantiatedby a study of the sensitivity properties of naive OBCs [5]. In thispaper we investigate the sensitivity of MBCs. We present sensitivityfunctions for the outcome probabilities of interest of an MBC and usethese functions to study the sensitivity value. This value captures thesensitivity of an output probability to small changes in a parameter.We compare MBCs to OBCs in this respect and conclude that anMBC will on average be even more robust to parameter changes thanan OBC.

U2 - 10.3233/978-1-61499-419-0-971

DO - 10.3233/978-1-61499-419-0-971

M3 - Conference contribution

SN - 978-1-61499-418-3

T3 - Frontiers in Artificial Intelligence and Applications

SP - 971

EP - 972

BT - Proceedings of the 21st European Conference on Artificial Intelligence

A2 - Schaub, Torsten

A2 - Friedrich, Gerhard

A2 - O'Sullivan, Barry

PB - IOS Press

T2 - ECAI 2014

Y2 - 18 August 2014 through 22 August 2014

ER -

Sensitivity of multi-dimensional Bayesian classifiers

Abstract

Publication series

Conference

Access to Document

Fingerprint

Cite this