Relationship extraction method fusing entity type characterization and relation characterization

The invention discloses a relation extraction method fusing entity type representation and relation representation, and belongs to the technical field of relation extraction. According to the method, a text-subject-object weak related semantic representation mechanism is designed, entity type information is introduced to replace entity word meaning information, and then dependence of an extraction model on subject-object semantic association is reduced; on the basis, abstract semantic information of the entity relationship is further modeled and fused with context semantic representation containing subject-object type information, semantic mapping of the entity relationship is generated, and a more accurate prediction effect of a subject-relationship-object triple is obtained. 本发明公开了一种融合实体类型表征与关系表征的关系抽取方法，属于关系抽取技术领域。本发明设计文本-主客体弱相关语义表征机制，通过引入实体类型信息替换实体词意信息，进而降低抽取模型对主体-客体语义关联的依赖；在上述基础上，本发明进一步建模实体关系的抽象语义信息，并与包含主客体类型信息的上下文语义表征进行融合，生成实体关系的语义映射，获得主体-关系-客体三元组更准确的预测效果。