摘要
为从决策表中获取最简决策规则,在研究粒计算理论的基础上,提出了一种基于粒计算的决策规则提取算法。该算法通过引入λ阶粒库的概念,利用粒计算的方法对由同一决策表所生成的粒库,从不同的角度和多个层次进行研究,尝试从较低阶的粒库中提取出尽可能多的具有一定规则覆盖度与置信度的简洁规则。最后通过实例验证了该算法的有效性。
To extract compact decision rules from decision table,an algorithm for decision rules extraction based on granular computing is proposed.In the algorithm,a concept of A-rank granular base is defined,and granular bases of different ranks for the same decision table are brought out,using the methods of granular computing as many as possible,to extract decision rules which satisfy certain coverage and confidence limit of rules.Lastly,the validity for the algorithm is proved by analyzing examples.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第25期132-134,共3页
Computer Engineering and Applications
基金
国家自然科学基金No.60873104
河南省高校新世纪优秀人才支持计划No.2006HANCET-19
河南省教育厅自然科学基金No.2008B520019~~
关键词
粒语言
粒计算
粒库
决策规则
granular language
granular computing
granular base
rules extraction
