Utilize Discourse Relations to Segment Document for Effective Summarization

Li Jiazheng, Muhammad Rafi

Research output: Chapter in Book/Published conference outputConference publication

Abstract

This paper proposes a clause-based extractive summarization algorithm by ranking and extracting semantic clauses from the original document. Discourse structure relation is useful for identifying semantically important parts of the source document. We segment the document into clauses and evaluate the importance of clauses based on semantic relations, and then, rank and extract them coarsely, and utilize graph rank to refine the extracted clauses. This way can create a more concise summary with more information and less redundancy. Research reach the following results: 1) compared with the other summarization algorithms on different granularity, the clausebased summarization achieves higher recall score; and, 2) different discourse relations have different importance.
Original languageEnglish
Title of host publicationProceedings - 15th International Conference on Semantics, Knowledge and Grids
Subtitle of host publicationOn Big Data, AI and Future Interconnection Environment, SKG 2019
EditorsHai Zhuge, Xiaoping Sun
PublisherIEEE
Pages12-15
Number of pages4
ISBN (Electronic)978-1-7281-5823-5
ISBN (Print)978-1-7281-5824-2
DOIs
Publication statusPublished - 23 Mar 2020
Event2019 15th International Conference on Semantics, Knowledge and Grids (SKG) - Guangzhou, China
Duration: 17 Sep 201918 Sep 2019

Publication series

NameProceedings - 15th International Conference on Semantics, Knowledge and Grids: On Big Data, AI and Future Interconnection Environment, SKG 2019

Conference

Conference2019 15th International Conference on Semantics, Knowledge and Grids (SKG)
Period17/09/1918/09/19

Keywords

  • Discourse structure
  • Semantic link network
  • Text summarization

Fingerprint

Dive into the research topics of 'Utilize Discourse Relations to Segment Document for Effective Summarization'. Together they form a unique fingerprint.

Cite this