Paper: A Comparative Study of Mixture Models for Automatic Topic Segmentation of Multiparty Dialogues

ACL ID I08-2133
Title A Comparative Study of Mixture Models for Automatic Topic Segmentation of Multiparty Dialogues
Venue International Joint Conference on Natural Language Processing
Session Main Conference
Year 2008
Authors

In this article we address the task of auto- matic text structuring into linear and non- overlapping thematic episodes at a coarse level of granularity. In particular, we deal with topic segmentation on multi-party meeting recording transcripts, which pose specific challenges for topic segmentation models. We present a comparative study of two probabilistic mixture models. Based on lexical features, we use these models in parallel in order to generate a low dimen- sional input representation for topic segmen- tation. Our experiments demonstrate that in this manner important information is cap- tured from the data through less features.