The CRFs-Based Chinese Open Entity Relation Extraction
Xiaoyang Wu, Bin Wu · 2017
With the rapid development of the Internet, massive Internet text data has brought new opportunities and challenges to the research of entity relation extraction. Open entity relation extraction overcomes the shortage of traditional methods, that relation types need to be predefined and plenty of training data need to be labeled in advance. A lot of work have been done for English Open ERE, and now the Chinese Open ERE is attracting more and more researchers and scholars. This paper presents a novel CRFs-Based Semi-Supervised Chinese Open Entity Relation Extraction method-C-COERE. C-COERE combines the CRFs and the syntax parse tree to extract the entity relation tuples. Then it improves the accuracy of the extracted results through a post-processing filter module. Meanwhile, it enhances the recall of the extracted results by exploiting the duality of patterns and tuples. Experiments on real dataset show that the C-COERE method has achieved good effect.