Preview

Description

We present a method for partitioning expository texts into coherent multi-paragraph units which reflect the subtopic structure of the texts. Using Chafe's Flow Model of discourse, we observe that subtopics are often expressed by the interaction of multiple simultaneous themes. We describe two fully-implemented algorithms that use only term repetition information to determine the extents of the subtopics. We show that the segments correspond well to human judgements of the major subtopic boundaries of thirteen lengthy texts, and suggest the use of such segments in information retrieval applications.

Details

Files

Actions

Statistics

from
to
Export
Download Full History