|
|||||
|
|
| 基于知网的汉语语义实例库的建设与应用 | |
| 引用本文： | 苗雪雷,蔡东风,苗洪霞. 基于知网的汉语语义实例库的建设与应用[J]. 沈阳航空工业学院学报, 2007, 24(1): 32-34 |
| 作者姓名： | 苗雪雷 蔡东风 苗洪霞 |
| 作者单位： | 沈阳航空工业学院自然语言处理研究室,辽宁,沈阳,110034 |
| 摘 要： | 为了解决有指导词义消歧任务中的知识源问题,提出并构建了一个基于知网的汉语词义实例库(CSIC),同时开发了一个语义标注平台(SenseTag)。该平台通过方便快捷的人机交互方式显著提高了实例库的建设效率和质量。利用条件随机场模型,从实例库中自动学习消歧知识,进行自动标准在随机选取的部分汉语高频多义词的词义消歧开放测试中,取得了平均正确率85%的较好效果。 |
| 关 键 词： | 汉语词义实例库 知网 词义消歧 条件随机场 |
| 文章编号： | 1007-1385（2007）01-0032-03 |
| 收稿时间： | 2006-10-15 |
| 修稿时间： | 2006-10-15 |
Construction and use of Chinese sense instance corpus on large scale HowNet |
|
| MIAO Xue-lei,CAI Dong-feng,Miao Hong-xia. Construction and use of Chinese sense instance corpus on large scale HowNet[J]. Journal of Shenyang Institute of Aeronautical Engineering, 2007, 24(1): 32-34 | |
| Authors: | MIAO Xue-lei CAI Dong-feng Miao Hong-xia |
| Affiliation: | Natural Language Processing Lab, Shenyang Institute of Aeronautical Engineering, Liaoning Shenyang, 110034 |
| Abstract: | This paper constructed a Chinese Sense Instance Corpus(CSIC) based on HowNet to solve the re-source of knowledge problem.At the same time,a tagging platform was developed to enhance the construction quality and to speed by quick and convenient human-machine mode.Testing result shows that the platform produces a marked effect.In the experiment using Cond itionalRandom Fields(CRF) model on some words token at random with higher frequencies and multi-senses,this method takes good performance in open testing environment,the accuracy is 85%. |
| Keywords: | Chinese sense instance corpus HowNet word sense disambiguation conditional random fields |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |