| 一种基于语义和统计特征的中文文本特征表示方法 |
| |
| 引用本文： | 赵鹏,耿焕同,蔡庆生. 一种基于语义和统计特征的中文文本特征表示方法[J]. 小型微型计算机系统, 2007, 28(7): 1311-1313 |
| |
| 作者姓名： | 赵鹏 耿焕同 蔡庆生 |
| |
| 作者单位： | 1. 中国科学技术大学,计算机系,安徽,合肥,230027;安徽大学,计算机系,安徽,合肥,230039
2. 中国科学技术大学,计算机系,安徽,合肥,230027 |
| |
| 基金项目： | 国家自然科学基金;安徽省自然科学基金;安徽省高校青年教师科研项目 |
| |
| 摘 要： | 基于关键词集的中文文本特征表示方法难以准确表示文本语义信息,从而导致聚类质量较差.为了解决这个问题,本文将本体论和词共现模型的思想引入到中文文本的特征表示中,并在此基础上提出了一种基于语义和统计特征的中文文本特征表示方法.本方法在统计特征的基础上加入了基于知网和特征项共现的语义特征,实验结果表明该方法更加准确地表示了中文文本的语义信息,使得中文文本自动聚类的质量提高了近18%.
|
| 关 键 词： | 向量空间模型 本体论 知网 词共现 |
| 文章编号： | 1000-1220（2007）07-1300-03 |
| 修稿时间： | 2006-02-21 |
An Approach of Chinese Text Representation Based on Semantic and Statistic Feature |
| |
| ZHAO Peng,GENG Huan-tong,CAI Qing-sheng. An Approach of Chinese Text Representation Based on Semantic and Statistic Feature[J]. Mini-micro Systems, 2007, 28(7): 1311-1313 |
| |
| Authors: | ZHAO Peng GENG Huan-tong CAI Qing-sheng |
| |
| Affiliation: | Departrnent of Computer Science and Technology, University of Science and Technology of China, Hefei 230027, China;2.Departrnent of Computer Science , Anhui University, Hefei 230039, China |
| |
| Abstract: | The approach of Chinese text representation based on keywords set cannot represent the semantic information of text, and then results in low quality of text clustering. To settle this problem, this paper introduces Ontology and the idea of Term Co-occurrence into Chinese text representation and presents an approach of Chinese text representation based on semantic and statistic feature. This approach adds semantic feature based on Hownet and feature co-occurrence. Experimental results show that this approach can represent the semantic information of text more precisely and improve the quality of text clustering greatly. |
| |
| Keywords: | vector space model ontology HowNet term co-occurrence |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
