Speech act profiling: A probabilistic method for analyzing persistent conversations and their participants

Douglas P. Twitchell, Jay F Nunamaker

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

16 Citations (Scopus)

Abstract

The increase in persistent conversations in the form of chat and instant messaging (IM) has presented new opportunities for researchers. This paper describes a method for evaluating and visualizing persistent conversations by creating a speech act profile for conversation participants using speech act theory and concepts from fuzzy logic. This method can be used either to score a participant based on possible intentions or to create a visual map of those intentions. Transcripts from the Switchboard corpus, which have been marked up with speech act labels according to a SWBD-DAMSL tag set of 42 tags, are used to train language models and a modified hidden Markov model (HMM) to obtain probabilities for each speech act type for a given sentence. Rather than choosing the speech act with the maximum probability and assigning it to the sentence, the probabilities are aggregated for each conversation participant creating a set of speech act profiles, which can be visualized as a radar graphs. Several example profiles are shown along with possible interpretations. The profiles can be used as an overall picture of a conversation, and may be useful in various analyses of persistent conversations including information retrieval, deception detection, and online technical support monitoring.

Original languageEnglish (US)
Title of host publicationProceedings of the Hawaii International Conference on System Sciences
EditorsR.H. Sprague Jr.
Pages1713-1722
Number of pages10
Volume37
StatePublished - 2004
EventProceedings of the Hawaii International Conference on System Sciences - Big Island, HI., United States
Duration: Jan 5 2004Jan 8 2004

Other

OtherProceedings of the Hawaii International Conference on System Sciences
CountryUnited States
CityBig Island, HI.
Period1/5/041/8/04

Fingerprint

Hidden Markov models
Information retrieval
Fuzzy logic
Labels
Radar
Monitoring

ASJC Scopus subject areas

  • Software
  • Industrial and Manufacturing Engineering

Cite this

Twitchell, D. P., & Nunamaker, J. F. (2004). Speech act profiling: A probabilistic method for analyzing persistent conversations and their participants. In R. H. Sprague Jr. (Ed.), Proceedings of the Hawaii International Conference on System Sciences (Vol. 37, pp. 1713-1722). [DDPCN06]

Speech act profiling : A probabilistic method for analyzing persistent conversations and their participants. / Twitchell, Douglas P.; Nunamaker, Jay F.

Proceedings of the Hawaii International Conference on System Sciences. ed. / R.H. Sprague Jr. Vol. 37 2004. p. 1713-1722 DDPCN06.

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

Twitchell, DP & Nunamaker, JF 2004, Speech act profiling: A probabilistic method for analyzing persistent conversations and their participants. in RH Sprague Jr. (ed.), Proceedings of the Hawaii International Conference on System Sciences. vol. 37, DDPCN06, pp. 1713-1722, Proceedings of the Hawaii International Conference on System Sciences, Big Island, HI., United States, 1/5/04.
Twitchell DP, Nunamaker JF. Speech act profiling: A probabilistic method for analyzing persistent conversations and their participants. In Sprague Jr. RH, editor, Proceedings of the Hawaii International Conference on System Sciences. Vol. 37. 2004. p. 1713-1722. DDPCN06
Twitchell, Douglas P. ; Nunamaker, Jay F. / Speech act profiling : A probabilistic method for analyzing persistent conversations and their participants. Proceedings of the Hawaii International Conference on System Sciences. editor / R.H. Sprague Jr. Vol. 37 2004. pp. 1713-1722
@inproceedings{e067f82f734d47ad9661dee3987fbaaa,
title = "Speech act profiling: A probabilistic method for analyzing persistent conversations and their participants",
abstract = "The increase in persistent conversations in the form of chat and instant messaging (IM) has presented new opportunities for researchers. This paper describes a method for evaluating and visualizing persistent conversations by creating a speech act profile for conversation participants using speech act theory and concepts from fuzzy logic. This method can be used either to score a participant based on possible intentions or to create a visual map of those intentions. Transcripts from the Switchboard corpus, which have been marked up with speech act labels according to a SWBD-DAMSL tag set of 42 tags, are used to train language models and a modified hidden Markov model (HMM) to obtain probabilities for each speech act type for a given sentence. Rather than choosing the speech act with the maximum probability and assigning it to the sentence, the probabilities are aggregated for each conversation participant creating a set of speech act profiles, which can be visualized as a radar graphs. Several example profiles are shown along with possible interpretations. The profiles can be used as an overall picture of a conversation, and may be useful in various analyses of persistent conversations including information retrieval, deception detection, and online technical support monitoring.",
author = "Twitchell, {Douglas P.} and Nunamaker, {Jay F}",
year = "2004",
language = "English (US)",
volume = "37",
pages = "1713--1722",
editor = "{Sprague Jr.}, R.H.",
booktitle = "Proceedings of the Hawaii International Conference on System Sciences",

}

TY - GEN

T1 - Speech act profiling

T2 - A probabilistic method for analyzing persistent conversations and their participants

AU - Twitchell, Douglas P.

AU - Nunamaker, Jay F

PY - 2004

Y1 - 2004

N2 - The increase in persistent conversations in the form of chat and instant messaging (IM) has presented new opportunities for researchers. This paper describes a method for evaluating and visualizing persistent conversations by creating a speech act profile for conversation participants using speech act theory and concepts from fuzzy logic. This method can be used either to score a participant based on possible intentions or to create a visual map of those intentions. Transcripts from the Switchboard corpus, which have been marked up with speech act labels according to a SWBD-DAMSL tag set of 42 tags, are used to train language models and a modified hidden Markov model (HMM) to obtain probabilities for each speech act type for a given sentence. Rather than choosing the speech act with the maximum probability and assigning it to the sentence, the probabilities are aggregated for each conversation participant creating a set of speech act profiles, which can be visualized as a radar graphs. Several example profiles are shown along with possible interpretations. The profiles can be used as an overall picture of a conversation, and may be useful in various analyses of persistent conversations including information retrieval, deception detection, and online technical support monitoring.

AB - The increase in persistent conversations in the form of chat and instant messaging (IM) has presented new opportunities for researchers. This paper describes a method for evaluating and visualizing persistent conversations by creating a speech act profile for conversation participants using speech act theory and concepts from fuzzy logic. This method can be used either to score a participant based on possible intentions or to create a visual map of those intentions. Transcripts from the Switchboard corpus, which have been marked up with speech act labels according to a SWBD-DAMSL tag set of 42 tags, are used to train language models and a modified hidden Markov model (HMM) to obtain probabilities for each speech act type for a given sentence. Rather than choosing the speech act with the maximum probability and assigning it to the sentence, the probabilities are aggregated for each conversation participant creating a set of speech act profiles, which can be visualized as a radar graphs. Several example profiles are shown along with possible interpretations. The profiles can be used as an overall picture of a conversation, and may be useful in various analyses of persistent conversations including information retrieval, deception detection, and online technical support monitoring.

UR - http://www.scopus.com/inward/record.url?scp=12344315829&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=12344315829&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:12344315829

VL - 37

SP - 1713

EP - 1722

BT - Proceedings of the Hawaii International Conference on System Sciences

A2 - Sprague Jr., R.H.

ER -