Proceedings Chapter (114 Kb) - Free access
Towards Automatic Identification of Discourse Markers in Dialogs: The Case of 'Like'
|Published in||Michael Strube and Candy Sidner. SIGdial 2004 (5th SIGdial Workshop on Discourse and Dialogue). Cambridge (Mass., USA): ACL - Association for Computational Linguistics. 2004, p. 63-71|
|Abstract||This article discusses the detection of discourse markers (DM) in dialog transcriptions, by human annotators and by automated means. After a theoretical discussion of the definition of DMs and their relevance to natural language processing, we focus on the role of like as a DM. Results from experiments with human annotators show that detection of DMs is a difficult but reliable task, which requires prosodic information from soundtracks. Then, several types of features are defined for automatic disambiguation of like: collocations, part-of-speech tags and duration-based features. Decision-tree learning shows that for like, nearly 70% precision can be reached, with near 100% recall, mainly using collocation filters. Similar results hold for well, with about 91% precision at 100% recall.|
|ZUFFEREY, Sandrine, POPESCU-BELIS, Andréi. Towards Automatic Identification of Discourse Markers in Dialogs: The Case of 'Like'. In: Michael Strube and Candy Sidner (Ed.). SIGdial 2004 (5th SIGdial Workshop on Discourse and Dialogue). Cambridge (Mass., USA). [s.l.] : ACL - Association for Computational Linguistics, 2004. p. 63-71. https://archive-ouverte.unige.ch/unige:2273|