Hello, who is calling? Canwords reveal the social nature of conversations?

Anthony Stark, Izhak Shafran, Jeffrey Kaye

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Citations (Scopus)

Abstract

This study aims to infer the social nature of conversations from their content automatically. To place this work in context, our motivation stems from the need to understand how social disengagement affects cognitive decline or depression among older adults. For this purpose, we collected a comprehensive and naturalistic corpus comprising of all the incoming and outgoing telephone calls from 10 subjects over the duration of a year. As a first step, we learned a binary classifier to filter out business related conversation, achieving an accuracy of about 85%. This classification task provides a convenient tool to probe the nature of telephone conversations. We evaluated the utility of openings and closing in differentiating personal calls, and find that empirical results on a large corpus do not support the hypotheses by Schegloff and Sacks that personal conversations are marked by unique closing structures. For classifying different types of social relationships such as family vs other, we investigated features related to language use (entropy), hand-crafted dictionary (LIWC) and topics learned using unsupervised latent Dirichlet models (LDA). Our results show that the posteriors over topics from LDA provide consistently higher accuracy (60-81%) compared to LIWC or language use features in distinguishing different types of conversations.

Original languageEnglish (US)
Title of host publicationNAACL HLT 2012 - 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference
PublisherAssociation for Computational Linguistics (ACL)
Pages112-119
Number of pages8
ISBN (Print)1937284204, 9781937284206
StatePublished - 2012
Event2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL HLT 2012 - Montreal, Canada
Duration: Jun 3 2012Jun 8 2012

Other

Other2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL HLT 2012
CountryCanada
CityMontreal
Period6/3/126/8/12

Fingerprint

Telephone
conversation
Glossaries
Classifiers
Entropy
telephone
disengagement
Industry
entropy
language
dictionary
Linear Discriminant Analysis
Language Use
Telephone Calls
Filter
Telephone Conversation
Dirichlet
Classifier
Dictionary
Disengagement

ASJC Scopus subject areas

  • Language and Linguistics
  • Computer Science Applications
  • Linguistics and Language

Cite this

Stark, A., Shafran, I., & Kaye, J. (2012). Hello, who is calling? Canwords reveal the social nature of conversations? In NAACL HLT 2012 - 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference (pp. 112-119). Association for Computational Linguistics (ACL).

Hello, who is calling? Canwords reveal the social nature of conversations? / Stark, Anthony; Shafran, Izhak; Kaye, Jeffrey.

NAACL HLT 2012 - 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference. Association for Computational Linguistics (ACL), 2012. p. 112-119.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Stark, A, Shafran, I & Kaye, J 2012, Hello, who is calling? Canwords reveal the social nature of conversations? in NAACL HLT 2012 - 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference. Association for Computational Linguistics (ACL), pp. 112-119, 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL HLT 2012, Montreal, Canada, 6/3/12.
Stark A, Shafran I, Kaye J. Hello, who is calling? Canwords reveal the social nature of conversations? In NAACL HLT 2012 - 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference. Association for Computational Linguistics (ACL). 2012. p. 112-119
Stark, Anthony ; Shafran, Izhak ; Kaye, Jeffrey. / Hello, who is calling? Canwords reveal the social nature of conversations?. NAACL HLT 2012 - 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference. Association for Computational Linguistics (ACL), 2012. pp. 112-119
@inproceedings{f600fa6861d14ce6b459d4e0e5fb6e42,
title = "Hello, who is calling?: Canwords reveal the social nature of conversations?",
abstract = "This study aims to infer the social nature of conversations from their content automatically. To place this work in context, our motivation stems from the need to understand how social disengagement affects cognitive decline or depression among older adults. For this purpose, we collected a comprehensive and naturalistic corpus comprising of all the incoming and outgoing telephone calls from 10 subjects over the duration of a year. As a first step, we learned a binary classifier to filter out business related conversation, achieving an accuracy of about 85{\%}. This classification task provides a convenient tool to probe the nature of telephone conversations. We evaluated the utility of openings and closing in differentiating personal calls, and find that empirical results on a large corpus do not support the hypotheses by Schegloff and Sacks that personal conversations are marked by unique closing structures. For classifying different types of social relationships such as family vs other, we investigated features related to language use (entropy), hand-crafted dictionary (LIWC) and topics learned using unsupervised latent Dirichlet models (LDA). Our results show that the posteriors over topics from LDA provide consistently higher accuracy (60-81{\%}) compared to LIWC or language use features in distinguishing different types of conversations.",
author = "Anthony Stark and Izhak Shafran and Jeffrey Kaye",
year = "2012",
language = "English (US)",
isbn = "1937284204",
pages = "112--119",
booktitle = "NAACL HLT 2012 - 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference",
publisher = "Association for Computational Linguistics (ACL)",

}

TY - GEN

T1 - Hello, who is calling?

T2 - Canwords reveal the social nature of conversations?

AU - Stark, Anthony

AU - Shafran, Izhak

AU - Kaye, Jeffrey

PY - 2012

Y1 - 2012

N2 - This study aims to infer the social nature of conversations from their content automatically. To place this work in context, our motivation stems from the need to understand how social disengagement affects cognitive decline or depression among older adults. For this purpose, we collected a comprehensive and naturalistic corpus comprising of all the incoming and outgoing telephone calls from 10 subjects over the duration of a year. As a first step, we learned a binary classifier to filter out business related conversation, achieving an accuracy of about 85%. This classification task provides a convenient tool to probe the nature of telephone conversations. We evaluated the utility of openings and closing in differentiating personal calls, and find that empirical results on a large corpus do not support the hypotheses by Schegloff and Sacks that personal conversations are marked by unique closing structures. For classifying different types of social relationships such as family vs other, we investigated features related to language use (entropy), hand-crafted dictionary (LIWC) and topics learned using unsupervised latent Dirichlet models (LDA). Our results show that the posteriors over topics from LDA provide consistently higher accuracy (60-81%) compared to LIWC or language use features in distinguishing different types of conversations.

AB - This study aims to infer the social nature of conversations from their content automatically. To place this work in context, our motivation stems from the need to understand how social disengagement affects cognitive decline or depression among older adults. For this purpose, we collected a comprehensive and naturalistic corpus comprising of all the incoming and outgoing telephone calls from 10 subjects over the duration of a year. As a first step, we learned a binary classifier to filter out business related conversation, achieving an accuracy of about 85%. This classification task provides a convenient tool to probe the nature of telephone conversations. We evaluated the utility of openings and closing in differentiating personal calls, and find that empirical results on a large corpus do not support the hypotheses by Schegloff and Sacks that personal conversations are marked by unique closing structures. For classifying different types of social relationships such as family vs other, we investigated features related to language use (entropy), hand-crafted dictionary (LIWC) and topics learned using unsupervised latent Dirichlet models (LDA). Our results show that the posteriors over topics from LDA provide consistently higher accuracy (60-81%) compared to LIWC or language use features in distinguishing different types of conversations.

UR - http://www.scopus.com/inward/record.url?scp=84896371714&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84896371714&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84896371714

SN - 1937284204

SN - 9781937284206

SP - 112

EP - 119

BT - NAACL HLT 2012 - 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference

PB - Association for Computational Linguistics (ACL)

ER -