基于信息增益的缺省规则的挖掘方法

Mining Default Rule Using Informtion Gain

Rough集方法是一种处理不确定或模糊知识的重要工具。论文在现有的基于Rough集理论的缺省规则挖掘算法的基础上,将单属性信息增益概念扩充为多属性的情况,提出了基于信息增益的缺省规则的搜索策略和挖掘方法。实验表明,该方法能够发现简洁、易理解和实用的规则,同时具有较低的计算复杂性。

The Rough set approach is an important tool to deal with uncertain or vague knowledge in AI applications. In this paper,the concept of multi-attribute information gain is introduced,then algorithms for mining default rules based on Rough set and information gain are proposed and applied to mine the default rule from the decision table. Experimental results show that this algorithm can effectively obtain concise,apprehensible and practicable default results.