Abstract

We present results on addressee identification in four-participants face-to-face meetings using Bayesian Network and Naive Bayes classifiers. First, we investigate how well the addressee of a dialogue act can be predicted based on gaze, utterance and conversational context features. Then, we explore whether information about meeting context can aid classifiers’ performances. Both classifiers perform the best when conversational context and utterance features are combined with speaker’s gaze information. The classifiers show little gain from information about meeting context.
Original languageUndefined
Title of host publicationProceedings of 11th Conference of the European Chapter of the ACL (EACL)
EditorsD. McCarthy, S. Wintner
Place of PublicationPennsylvania, USA
PublisherAssociation for Computational Linguistics
Pages169-176
Number of pages8
ISBN (Print)1-932432-59-0
StatePublished - Apr 2006

Publication series

Name
PublisherAssociation for Computational Linguistics
Numbersuppl 2

Fingerprint

Classifiers
Bayesian networks

Keywords

  • HMI-CI: Computational Intelligence
  • EWI-8802
  • EC Grant Agreement nr.: FP6/506811
  • IR-66786
  • METIS-237820

Cite this

Jovanovic, N., op den Akker, H. J. A., & Nijholt, A. (2006). Addressee Identification In Face-to-Face Meetings. In D. McCarthy, & S. Wintner (Eds.), Proceedings of 11th Conference of the European Chapter of the ACL (EACL) (pp. 169-176). Pennsylvania, USA: Association for Computational Linguistics.

Jovanovic, N.; op den Akker, Hendrikus J.A.; Nijholt, Antinus / Addressee Identification In Face-to-Face Meetings.

Proceedings of 11th Conference of the European Chapter of the ACL (EACL). ed. / D. McCarthy; S. Wintner. Pennsylvania, USA : Association for Computational Linguistics, 2006. p. 169-176.

Research output: Scientific - peer-reviewConference contribution

@inbook{dc58faf119cf4123be2fc96bfe665015,
title = "Addressee Identification In Face-to-Face Meetings",
abstract = "We present results on addressee identification in four-participants face-to-face meetings using Bayesian Network and Naive Bayes classifiers. First, we investigate how well the addressee of a dialogue act can be predicted based on gaze, utterance and conversational context features. Then, we explore whether information about meeting context can aid classifiers’ performances. Both classifiers perform the best when conversational context and utterance features are combined with speaker’s gaze information. The classifiers show little gain from information about meeting context.",
keywords = "HMI-CI: Computational Intelligence, EWI-8802, EC Grant Agreement nr.: FP6/506811, IR-66786, METIS-237820",
author = "N. Jovanovic and {op den Akker}, {Hendrikus J.A.} and Antinus Nijholt",
year = "2006",
month = "4",
isbn = "1-932432-59-0",
publisher = "Association for Computational Linguistics",
number = "suppl 2",
pages = "169--176",
editor = "D. McCarthy and S. Wintner",
booktitle = "Proceedings of 11th Conference of the European Chapter of the ACL (EACL)",

}

Jovanovic, N, op den Akker, HJA & Nijholt, A 2006, Addressee Identification In Face-to-Face Meetings. in D McCarthy & S Wintner (eds), Proceedings of 11th Conference of the European Chapter of the ACL (EACL). Association for Computational Linguistics, Pennsylvania, USA, pp. 169-176.

Addressee Identification In Face-to-Face Meetings. / Jovanovic, N.; op den Akker, Hendrikus J.A.; Nijholt, Antinus.

Proceedings of 11th Conference of the European Chapter of the ACL (EACL). ed. / D. McCarthy; S. Wintner. Pennsylvania, USA : Association for Computational Linguistics, 2006. p. 169-176.

Research output: Scientific - peer-reviewConference contribution

TY - CHAP

T1 - Addressee Identification In Face-to-Face Meetings

AU - Jovanovic,N.

AU - op den Akker,Hendrikus J.A.

AU - Nijholt,Antinus

PY - 2006/4

Y1 - 2006/4

N2 - We present results on addressee identification in four-participants face-to-face meetings using Bayesian Network and Naive Bayes classifiers. First, we investigate how well the addressee of a dialogue act can be predicted based on gaze, utterance and conversational context features. Then, we explore whether information about meeting context can aid classifiers’ performances. Both classifiers perform the best when conversational context and utterance features are combined with speaker’s gaze information. The classifiers show little gain from information about meeting context.

AB - We present results on addressee identification in four-participants face-to-face meetings using Bayesian Network and Naive Bayes classifiers. First, we investigate how well the addressee of a dialogue act can be predicted based on gaze, utterance and conversational context features. Then, we explore whether information about meeting context can aid classifiers’ performances. Both classifiers perform the best when conversational context and utterance features are combined with speaker’s gaze information. The classifiers show little gain from information about meeting context.

KW - HMI-CI: Computational Intelligence

KW - EWI-8802

KW - EC Grant Agreement nr.: FP6/506811

KW - IR-66786

KW - METIS-237820

M3 - Conference contribution

SN - 1-932432-59-0

SP - 169

EP - 176

BT - Proceedings of 11th Conference of the European Chapter of the ACL (EACL)

PB - Association for Computational Linguistics

ER -

Jovanovic N, op den Akker HJA, Nijholt A. Addressee Identification In Face-to-Face Meetings. In McCarthy D, Wintner S, editors, Proceedings of 11th Conference of the European Chapter of the ACL (EACL). Pennsylvania, USA: Association for Computational Linguistics. 2006. p. 169-176.