| 摘要: |
| 进化式信息过滤方法从多个角度描述用户的信息需求,通过类似自然选择的过程,达到系统整体过滤性能的优化.同时还从面向对象程序设计语言的设计思想中获得启发,引入了继承、类树的概念,增加了过滤系统的易用性.这种方法可以缩短训练时间，提高过滤质量,同时减小过滤结果与训练次序的相关性. |
| 关键词: 信息过滤,信息处理. |
| DOI： |
| 分类号: |
| 基金项目:本文研究得到国家863高科技项目基金(No.863-306-zt01-03-1)和IBM另研究中心基金资助. |
|
| Evolving Information Filtering Method |
|
TIAN Fan-jiang,LI Cong-rong,WANG Ding-xing
|
| Abstract: |
| Evolving information filtering method presented in this paper describes user's information needs from multi-aspects simultaneously,and improves system filtering performance through a process like natural selection.Inspired by object-oriented programming language,this method also introduces concepts such as inherit and classtree,which make filtering systems easy to use.This method can shorten training time,improve filtering precision,and reduce the relevance between filtering results and training sequence. |
| Key words: Information filtering,information processing. |
