Open Domain Event Attribute Extraction Method

Junlei Zhao, Bicheng Li, Lei Guo

Abstract


To improve the extraction of attribute information related to open domain events, a method using Semantic Dependency Parsing (SDP) and a rule model is proposed. First, a small number of items are extracted from the corpus, and the template library corresponding to the attribute information is summarized. SDP is then used to process the entire corpus, identify unrecognized attribute information templates, and add them to the template library. Finally, the error-driven modification method is used to adjust the generated template library. Experimental results demonstrate that this method can enhance the performance of event attribute information extraction.


DOI
10.12783/dtcse/iciti2018/29124

Full Text:

PDF

Refbacks

  • There are currently no refbacks.