地质领域开放式实体关系联合抽取  

Joint extraction of open entity relation in geological field

作　　者：王智广[1,2] 文红英 鲁强[1,2] 申昊锴 WANG Zhi-guang;WEN Hong-ying;LU Qiang;SHEN Hao-kai(Department of Computer Science and Technology,China University of Petroleum-Beijing,Beijing 102249,China;Beijing Key Laboratory of Petroleum Data Mining,China University of Petroleum-Beijing,Beijing 102249,China)

机构地区：[1]中国石油大学(北京)计算机科学与技术系,北京102249 [2]中国石油大学(北京)石油数据挖掘北京市重点实验室,北京102249

出　　处：《计算机工程与设计》2021年第4期996-1005,共10页Computer Engineering and Design

基　　金：国家重点研发计划基金项目(2019YFC0312003);国家科技重大基金项目(2017ZX05018-005)。

摘　　要：为解决地质领域实体关系类型复杂且缺乏大量标注语料的问题,建立基于句法结构的开放式实体关系联合抽取模型CSSEM(Chinese syntactic structure extraction model)。给出一种基于模式的地质领域实体识别方法,解决地质领域缺乏实体标注语料的问题;基于少量标注语料自动学习关系抽取模式,使用基于句法结构的抽取模式从非结构化文本中开放式地抽取三元组。基于AUC(area under curve)评价标准,CSSEM与其它方法在地质领域和通用领域的数据集上进行对比分析。分析结果表明,该模型在上述两个领域取得了更好的准确率和召回率。To solve the problems of complex types of entities and relations as well as the lack of a large number of annotated corpora in the geological field,an open model CSSEM(Chinese syntactic structure extraction model)of entity relation extraction was established,in which relation was extracted based on syntactic structure.A pattern-based method was proposed for identifying entities in the geological field,which solved the problem of lack of entity annotation corpus.Relation extraction patterns were automatically learned based on a small amount of annotated corpora,and syntactic structure-based patterns were used to openly extract triples from unstructured text.Based on the AUC(area under curve)evaluation standard,CSSEM and other methods were compared and analyzed in the geological and general dataset.The results show that the model achieves better precision and recall in the field of geology and general.

关 键 词：地质领域 句法结构 依存句法分析 实体关系抽取 模式 中文句法结构抽取模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]