作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2012, Vol. 38 ›› Issue (21): 197-201. doi: 10.3969/j.issn.1000-3428.2012.21.053

• 人工智能及识别技术 • 上一篇    下一篇

基于近似随机测试的语义关系抽取比较

彭 成a,钱龙华a,b,赵知纬a,周国栋a,b   

  1. (苏州大学 a. 计算机科学与技术学院;b. 自然语言处理实验室,江苏 苏州 215006)
  • 收稿日期:2011-12-26 出版日期:2012-11-05 发布日期:2012-11-02
  • 作者简介:彭 成(1987-),男,硕士研究生、CCF会员,主研方向:信息抽取;钱龙华,副教授;赵知纬,硕士研究生;周国栋,教授、博士生导师
  • 基金资助:
    国家自然科学基金资助项目(60873150, 90920004, 61003153);江苏省自然科学基金资助项目(BK2010219)

Comparison on Semantic Relation Extraction Based on Approximation Random Test

PENG Cheng a, QIAN Long-hua a,b, ZHAO Zhi-wei a, ZHOU Guo-dong a,b   

  1. (a. School of Computer Science & Technology; b. Natural Language Processing Lab, Soochow University, Suzhou 215006, China)
  • Received:2011-12-26 Online:2012-11-05 Published:2012-11-02

摘要: 为比较结构化信息和句法分析器对树核函数的关系抽取的作用,提出一种基于近似随机测试语义关系比较方法。对于2种不同配置关系的抽取结果,采用随机标号互换的方法重复产生样本,通过计算这些样本的性能差异进行显著性分析。实验结果表明,动态关系树是最佳的结构化信息,句法分析器Charniak和Berkeley性能均优于Stanford。

关键词: 关系抽取, 树核函数, 结构化信息, 显著性测试, 近似随机测试

Abstract: To scientifically compare the effect of structured information and parsers on kernel-based relation extraction, a comparison method based on random approximate test is proposed. It gives two relation extraction results for different settings, samples are produced repeatedly using random label exchange from re-sampling techniques, and significant tests are conducted by calculating the performance differences between these samples. Experimental results show that dynamic relation tree is the best structured information, and the performance of Charniak and Berkeley are better than Stanford.

Key words: relation extraction, tree kernel function, structured information, significance test, approximate random test

中图分类号: