A lexically-driven algorithm for disfluency detection

TitleA lexically-driven algorithm for disfluency detection
Publication TypeConference Papers
Year of Publication2004
AuthorsSnover M, Dorr BJ, Schwartz R
Conference NameProceedings of HLT-NAACL 2004: Short Papers
Date Published2004///
PublisherAssociation for Computational Linguistics
Conference LocationStroudsburg, PA, USA
ISBN Number1-932432-24-8

This paper describes a transformation-based learning approach to disfluency detection in speech transcripts using primarily lexical features. Our method produces comparable results to two other systems that make heavy use of prosodic features, thus demonstrating that reasonable performance can be achieved without extensive prosodic cues. In addition, we show that it is possible to facilitate the identification of less frequently disfluent discourse markers by taking speaker style into account.