基于统计的中文机构名实体识别的研究

针对中文机构名实体的识别提出了利用隐马尔科夫统计模型的方法.同时,在此方法的基础上通过构建一元模型来补充识别那些缩写的和没有核心词的机构名,弥补了单一统计模型在机构名实体识别中的不足.实验证明二者相结合的方法能获得较高的准确率和召回率,具有可行性.

The research about the Organization Entity Recognition based on statistic

This article put forward the method of the Chinese Organization Entity Recognition using statistic model based on the Hidden Markov Model. At the same time, those Organization Entities which were abbreviawry and had no nuclear vocable through designing Single Model can be also recognized based on this method. It can make up the limitation of single statistic model. The experiments proved that this combinative method gains higher precision and recall, and it is feasible.