| 对象迁移自动机在TDT中的聚类研究 |
| |
| 引用本文： | 施凡,陆余良,刘金红,夏阳.对象迁移自动机在TDT中的聚类研究[J].安徽大学学报(自然科学版),2007,31(1):27-30. |
| |
| 作者姓名： | 施凡 陆余良 刘金红 夏阳 |
| |
| 作者单位： | 解放军电子工程学院,网络工程系,安徽,合肥,230037 |
| |
| 摘 要： | 对象迁移自动机(OMA)是一种能够较好地解决话题识别与跟踪(TDT)中聚类问题的方法,但是,传统OMA模型由于聚类速度慢等缺点,难以满足TDT实时和增量聚类的要求.针对这一问题,本文一方面改进传统的OMA模型中自动机的动作设计,同时提出文档选择策略,加快了OMA的聚类速度.改进的方法在中等文档集上进行了实验,实验结果表明,该方法具有较好的聚类效果.
|
| 关 键 词： | 对象迁移自动机 话题识别与跟踪 文本聚类 文本选择策略 |
| 文章编号： | 1000-2162（2007）01-0027-04 |
| 修稿时间： | 2006-09-20 |
A cluster reasearch of TDT based on object migration automaton |
| |
| SHI Fan,LU Yu-liang,LIU Jin-hong,XIA Yang.A cluster reasearch of TDT based on object migration automaton[J].Journal of Anhui University(Natural Sciences),2007,31(1):27-30. |
| |
| Authors: | SHI Fan LU Yu-liang LIU Jin-hong XIA Yang |
| |
| Institution: | Department of Network Engineering, Electronic Engineering Institute, Hefei 230037,China |
| |
| Abstract: | Object Migration Automaton(OMA) is a good method for Topic Detection and Tracking(TDT),because of its slow clustering,traditional OMA model can't satisfy the requirement in the aspect of real-time and incremental clustering.Therefore,this paper improves the traditional OMA by improved the actions and proposed the document-choosing policy of the automaton to accelerate the clustering of OMA.The improved method was experimented on a middle size document collection,the result indicates that the method has a better clustering performance. |
| |
| Keywords: | object migration automaton topic detection and tracking document clustering document-choosing policy |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
