| 摘要: |
| 本文在对ＣＯＢＷＥＢ、ＣＬＡＳＳＩＴ等概念聚类系统研究的基础上，提出了一种用数值属性的聚类分划来表示数值属性的方法．这种表示的核心是基于数值属性的取值分布．对于在这种表示下数值属性聚类的相关问题及性质，本文进行了较为详细的讨论．在此基础上，本文给出了一个能统一处理数值属性和符号属性的聚类评价函数．一个基于聚类分划表示方法的概念形成学习系统ＣＦＬＳ（ｃｏｎｃｅｐｔｆｏｒｍａｔｉｏｎｌｅａｒｎｉｎｇｓｙｓｔｅｍ）已在微机上实现，并被应用于地质学领域的三叶虫分类问题．本文对ＣＦＬＳ的设计和实现进行了介绍. |
| 关键词: 人工智能 知识获取 机器学习 学习系统 概念形成 |
| DOI： |
| 分类号: |
| 基金项目:本文研究得到地矿部高科技项目“矿产资源与地质环境专家系统”基金资助. |
|
| THE STUDY OF VHDL BASED SYSTEM SPECIFICATION LANGUAGES |
|
ZHANG Huajie,QIANG Fangzuo,YUAN Guobin
|
| Abstract: |
| Through the research on COBWEB and CLASSIT，this paper presents a new representation method of numeric attributes．This method represents numeric attributes based on its value distribution．The relative problems and properties of clustering in this representation are discussed in detail．This paper also proposes a clustering evaluation function which can deal with both symbolic attributes and numeric attributes uniformly．A concept formation system CFLS(concept formation learning system)has been implemented and applied to geological classification area．CFLS is also introduced in this paper． |
| Key words: Artificial intelligence knowledge acquisition machine learning learning systems concept formation |
