Detecting Uncertainty in Spoken Dialogues: An explorative research for the automatic detection of speaker uncertainty by using prosodic markers

Jeroen Dral, Dirk K.J. Heylen, Hendrikus J.A. op den Akker

    Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

    Abstract

    This paper reports results in automatic detection of speaker uncertainty in spoken dialogues by using prosodic markers. For this purpose a substantial part of the AMI corpus (a multi-modal multi-party meeting corpus) has been selected and converted to a suitable format so its data could be analyzed for a selected set of prosodic features. In the absence of relevant stance annotations on (un)certainty, lexical markers (hedges) have been used to mark utterances as (un)certain. Results show that prosodic features can indeed be used to detect speaker uncertainty in spoken dialogues. The classifiers can tell uncertain from neutral utterances with an accuracy of 75% which is 25% over the baseline.
    Original languageUndefined
    Title of host publicationAffective Computing and Sentiment Analysis
    EditorsKurshid Ahmad
    Place of PublicationLondon
    PublisherSpringer
    Pages67-77
    Number of pages11
    ISBN (Print)978-94-007-1756-5
    DOIs
    Publication statusPublished - 2011

    Publication series

    NameText, Speech, and Language Analysis
    PublisherSpringer Verlag
    Number45
    Volume45
    ISSN (Print)1386-291X

    Keywords

    • METIS-285103
    • IR-79612
    • EWI-21454
    • sentiment analysis prosody uncertainty
    • HMI-MR: MULTIMEDIA RETRIEVAL
    • EC Grant Agreement nr.: FP6/033812

    Cite this