摘要
在范例推理系统中 ,系统的学习会使范例库逐渐增大 ,一般来说范例库越大 ,知识越丰富 ,但也不能无限增加 ,否则会大大增加相似范例检索的时间 ,降低系统的总体性能 .因此范例推理学习系统必须有维护功能 ,主要目的是限制范例库的无限膨胀 ,且能保持系统的性能 .本文在给出一个改进的删除策略维护方案的同时 ,并从另一角度出发 ,提出一个基于范例增加的维护策略 ,以保证系统的性能不受影响 。
In case-based system, learning makes the case base expand quickly, which enable the search time becoming longer and longer. At this time ,the system needs maintenance. In general, there are some methods about maintenance ,for instance ,utility policy, deletion policy. This paper discusses emphatically a method based on case-categorized deletion policy ,based on this method we propose an improved deletion method . Meanwhile we put forward a new approach –based on addition policy. The experimental result shows that these methods have better performance than other methods.
出处
《小型微型计算机系统》
CSCD
北大核心
2003年第10期1825-1828,共4页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目 (60 0 75 0 15
60 2 73 40 3 )资助
安徽省教育厅自然科学基金项目 (2 0 0 1kj0 0 2 )资助
关键词
范例推理
范例库维护
删除
增加
case-based reasoning
case base maintenance
deletion
addition
