一个统计与规则相结合的中文命名实体识别系统  

Chinese named entity recognition system using statistics-based and rules-based method

作　　者：向晓雯[1] 史晓东[1] 曾华琳[1] 

机构地区：[1]厦门大学信息科学与技术学院,福建厦门361005

出　　处：《计算机应用》2005年第10期2404-2406,共3页journal of Computer Applications

基　　金：国家863计划项目(2002AA117010)

摘　　要：介绍了一个中文命名实体识别系统,该系统采用了统计与规则相结合的方法。整个识别过程主要分成两个步骤,首先使用隐马尔可夫模型进行词性标注,然后利用具有优先级别的匹配规则对第一步的结果进行修正和转换。同时,系统还对上下文相关的命名实体识别作了初步的尝试。在863组织的命名实体识别评测中,系统的准确率、召回率和F值分别达到了81.93%,78.20%,80.02%。This paper presented a Chinese named entity recognition system that combined the statistics-based and rulesbased method, The whole process was divided into two steps. First to use the hidden Markov model for part-of-speech tagging, and then made use of match rules to amend and convert the result of the HMM step. The system also made an attempt at context-sensitive Chinese named entity recognition. In a named entity test organized by the 863 program, the precision, recall and F-score of the system reach 81.93%, 78.20% and 80.02% respectively.

关 键 词：命名实体 隐马尔可夫模型 匹配规则 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]