基于粗糙集理论的知识库分解与规则获取

粗糙集理论可以通过对数据的约简从例子中学习,获取决策规则。但是,当知识库规模较大、条件属性个数较多时,存在提取规则速度慢、规则长度长等缺点。本文介绍了粗糙集理论的基本概念,提出了一种基于粗糙集的知识库分解算法。首先引入决策属性支持度的概念,在此基础上定义了一个属性选择量度,选择最佳目标属性对知识库进行分解,直到所有对象都被精确分类,从而得到具有一定支持度的规则集,得到的规则长度短,提取速度快。通过对一个简单实例的分析,证明了该算法的可行性。

Decomposition Of Knowledge Base and Rules Acquisition Based on Rough Set Theory

Rough set can obtain decision rule by learning from example, But it is not efficiently to the large knowledge base. To extract rules from large knowledge base efficiently and promptly , a novel knowledge base decomposition approach was proposed. First ,a feature selection measure was defined based on the conception of decision attribute support degree, and the knowledge base was decomposed with the feature selection measure. The decomposition went on until all of the objects were classed accurately. Finally, a rule set with certain support degree was obtained, and the length of rules was decreased. The experimental result shows that the algorithm is feasible.