作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

›› 2004, Vol. 30 ›› Issue (S1): 381-.

• 论文 • 上一篇    下一篇

一个中文新词识别特征的研究

秦浩伟,步丰林   

  1. 上海交通大学计算机科学与工程系,上海交通大学计算机科学与工程系 上海 200030
  • 发布日期:2004-12-29

Research on a Feature of Chinese New Word Identification

QIN Haowei,BU Fenglin   

  1. (Department of Computer Science and Engineering,Shanghai Jiaotong University,Shanghai 200030)
  • Published:2004-12-29

摘要: 未登录词识别是汉语自动分词的难题之一。中文新词多为时间敏感词或者领域相关词汇,占据未登录词很大比例。针对新词识别问 题,首先总结了已有新词识别特征,然后引入了一个语言学特征。实验结果表明该特征可以提高新词识别率。

关键词: 中文分词, 新词识别, 支持向量机, 语素生产率

Abstract: One of the problems of Chinese word segmentation is the identification of OOV (out of vocabulary) words. Chinese new words are often time-sensitive or domain specific, which take large percentage of OOV words. As to new words identification (NWI) problem, this paper summarizes the previous work on NWI, and then introduces a novel feature. Final result shows that the performance of NWl can be improved by the feature.

中图分类号: