基于粗集理论的决策规则约简算法的研究与应用

Research and Application of Algorithm for Decision Rule Reduction Based on Rough Set Theory

粗集理论是一种新型的数据挖掘方法 ,其中约简是它的重要概念 .粗集理论的主要思想是在保持分类能力不变的前提下 ,通过属性约简和值约简 ,导出问题的决策或分类规则 .本文不同于目前将属性约简作为决策规则挖掘的重点 ,而是在属性约简的基础上又进一步提出了值约简算法 ,大大提高了规则的挖掘效率 .该算法对原有算法进行了改进 ,最终可获得更简的决策规则 .实例分析表明了该算法是较优的。

Rough set theory is a new method in data mining, and reduction is its kernel content. The primary idea in rough set theory is to deduct the rule of decision or classification through reducing the attributes and keeping the ability of classification. Unlike the fact that attribute reduction is acted as an important content during decision rules mining now, value reduction is further presented based on attribute reduction and mining efficiency for rules is greatly improved. This algorithm is advanced after an original algorithm is improved. It can receive much simpler decision rules. The example results show that this algorithm is better and applied.