The Chinese Open Relation Extraction Based on Dependency Parsing
Yuzhao Wang, Yunfei Yang, Ruixue Zhao · 2017
Open relation extraction is one of the most promising relation extractioins.However,the precision and recall of the existing methods are far from satisfication especially for extraction from chinese text.In this paper,we propose a method based on dependency parsing, and develope a new system called C-Reverb.Specifically,we design regular expressions for relation phrase through analyzing the Chinese grammar rules.Based on statistics of a large amount of sentences ,we find the distribution between entities and relations to formulate the rules for entity identification.Afterwards, the extracted triples which lack of entities are simply filtered.The experimental results show that our method efficiently improves the precision and recall rate of Chineses relation extraction.Furthermore, relation expressions extracted by our method are more informative than the previous method.