Paper: A Lexically-Driven Algorithm For Disfluency Detection

ACL ID N04-4040
Title A Lexically-Driven Algorithm For Disfluency Detection
Venue Human Language Technologies
Session Short Paper
Year 2004
Authors

This paper describes a transformation-based learn- ing approach to disfluency detection in speech tran- scripts 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 addi- tion, we show that it is possible to facilitate the iden- tification of less frequently disfluent discourse mark- ers by taking speaker style into account.