We present a new method for discovering a segmental discourse structure of a document while categorizing segment function. We demonstrate how retrieval of noun phrases and pronominal forms, along with a zero-sum weighting scheme, determines topicalized segmentation. Futhermore, we use term distribution to aid in identifying the role that the segment performs in the document. Finally, we present results of evaluation in terms of precision and recall which surpass earlier approaches.
@article{arxiv.cs/9809020,
title = {Linear Segmentation and Segment Significance},
author = {Min-Yen Kan and Judith L. Klavans and Kathleen R. McKeown},
journal= {arXiv preprint arXiv:cs/9809020},
year = {2007}
}
Comments
9 pages, US Letter, 4 figures. Software License can be found at http://www.cs.columbia.edu/nlp/licenses/segmenterLicenseDownload.html