摘要
介绍命名实体识别在文本信息处理领域的重要地位,分析了中文命名实体识别存在的困难,介绍中文命名实体识别的一般过程、评价标准及方法。提出了一种在构造内部规则和外部规则的同时采用概率统计的中文命名实体的识别方法,并利用这种基于规则和统计相结合的方法。实验证明该方法获得了较高的准确率和召回率,具有可行性和合理性,同时也指出了它的局限性。
This paper described the important position of named entity recognition in the field of text information processing, analyzed the problem of named entity recognition, described the general process, the evaluation criteria and the methods of Chinese named entity recognition. This article puts forward the recognition method of the Chinese named entity through constructing internal and external rules and adopting the statistic method. The experiments proved that this method gains higher precision and recall and has the feasibility and rationality through recognizing and testing the datum. At the same time, the limitation of this method was analvzed.
出处
《情报科学》
CSSCI
北大核心
2012年第5期708-712,786,共6页
Information Science
基金
国家自然科学基金资助项目(60975034)
安徽省自然科学青年基金项目(10040606Q64)
安徽省高校省级自然科学基金(KJ2012Z401)
宿州学院科研开放平台项目(2011YKF10)
关键词
命名实体
文本特征
中文命名实体
识别
named entity
text feature
Chinese named entity
recognition
