中文人称代词指代消解的研究*

指代消解是文本信息处理中的一项重要任务,针对这一任务过于复杂，本文分析了中文突发事件语料中人称代词的特点, 提出一种基于语料库,运用规则预处理与最大熵模型相结合的人称代词消解方法。在ACE05 bnews中文测试语料上，分别与仅用基于规则和仅用统计方法进行了对比实验，结果表明该方法分别在召回率、准确率和F值上有一定的提高，本方法是中文人称代词指代消解的较好平台。

Research on personal pronoun anaphora resolution in Chinese

Anaphora resolution plays an important role in text information processing. For this task is too complex, after analyzing the Chinese personal pronouns emergency features of the corpus, the paper presented a new model for personal pronouns anaphora resolution based on corpus, which using rule pretreatment combined with maximum entropy. On ACE05 bnews corpus, compared with the method that based on rule or maximum entropy, the contrast experiment shows that the algorithm presented in the paper has better precision, recall and F value, and the model is a good system for Chinese personal pronoun anaphora resolution research.