Revisiting Cross-document Structure Theory for multi-document discourse parsing

作者：

Highlights：

• CST was refined by formalizing, pruning and organizing relations.

• Refinements improved the annotation agreement in CSTNews corpus.

• A parser was built based on the refined version of CST.

• The results obtained by the parser outperformed previous CST based parsers.

摘要

•CST was refined by formalizing, pruning and organizing relations.•Refinements improved the annotation agreement in CSTNews corpus.•A parser was built based on the refined version of CST.•The results obtained by the parser outperformed previous CST based parsers.

论文关键词：Discourse parsing,Multi-document processing,Cross-document Structure Theory,Machine learning

论文评审过程：Received 8 May 2013, Revised 19 December 2013, Accepted 29 December 2013, Available online 23 January 2014.

论文官网地址：https://doi.org/10.1016/j.ipm.2013.12.003