利用未标注语料改进实体名识别性能  

Named Entity Recognition with Un-annotated Data

作　　者：陈宁昱[1] 周雅倩[1] 黄萱菁[1] 吴立德[1] 

机构地区：[1]复旦大学计算机科学与工程系,上海200433

出　　处：《中文信息学报》2005年第2期7-11,27,共6页Journal of Chinese Information Processing

基　　金：国家自然科学基金资助项目 (6 0 10 30 14 ) ;上海市科委重点研究项目资助 (0 35 115 0 2 8)

摘　　要：本文主要介绍了一个利用最大熵进行实体名识别的系统以及所采用的模型和选取的特征。这些特征包括单词本身的词法词态特征和上下文信息。利用这些在任何语言的文本上都极易获得的特征 ,我们采用最大熵分类器构建了一个基准系统。在此基础上 ,我们首先通过网络资源建立了实体名词典知识库 ;并利用词典和基准系统在未标注语料上抽取出现的实体名作为辅助的训练语料 ;最后再将这些语料加入训练。实验结果表明 ,辅助的训练语料能够在一定程度上提高系统的性能。In this paper, we describe a system that applies maximum entropy (ME) models to the task of named entity recognition (NER). We start with an annotated corpus and a set of features. These features include the morphological features and context information. They are easily obtainable for almost any language. We build a baseline NE recognizer based on these features. We first construct a named entity thesaurus based on the information on the web. Then the baseline together with the thesaurus is used to extract the named entities and their context information from additional un annotated data. In turn, these data are incorporated into the final recognizer. The experiments prove that these data could further improve the recognition accuracy.

关 键 词：计算机应用 中文信息处理 实体名识别 最大熵 未标注语料 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]