【作者】 齐振宇； 赵军； 杨帆；

【Author】 Zhenyu Qi,Jun Zhao,Fan Yang National Laboratory of Pattern Recognition,Institute of Automation,Chinese Academy of Sciences Beijing 100190

【机构】 中国科学院自动化研究所模式识别国家重点试验室；

【摘要】 本文研究开放式中文命名实体识别问题。针对传统命名实体识别方法具有训练语料标注困难,白适应能力不强等缺点,研究人员提出了开放式的命名实体识别。但由于汉语的特殊性,使得开放式中文命名实体识别的相关研究很少,效果也不理想。本文提出了一种开放式中文命名实体识别方法,模拟语言习得的过程,从句子结构出发,通过分类器得到含有命名实体的碎片序列,进而从中抽取出命名实体。实验结果表明,该方法可以较好地从不限领域的中文文本中抽取不限类别的命名实体。更多还原

【Abstract】 This paper focused on the task of open named entity recognition of Chinese.The traditional methods have the following two main problems.Firstly,to label the corpse is very boring work.Secondly,these methods come to big problem when facing domain transfer.Aiming at these two problems,this paper proposed a new Chinese Named Entity Recognition method based on the simulation of language learning.This system first got a set of word sequences of NE or with NE in it.Then it extracted NEs from the set.The experiments show this method can perform well in NE recognition in open Chinese texts.更多还原