一个新的最优简单规则挖掘算法

New Algorithm for Optimal Minimal Rules Mining

关联规则的提取是知识发现和数据挖掘领域中的重要问题,粗集理论是研究规则挖掘的重要的数学工具.本文研究信息系统中最优简单关联规则挖掘算法.本文利用信息关联矩阵中元素特性,对其进行变换,直接从中发现关联规则潜在的条件元,以此作为规则挖掘算法的基础.本文的算法简单直观,能挖掘出信息系统中所有最优简单规则,而且有效地避免了通常属性约简过程中的NP-hard问题.本文以一个实例证明本方法的有效性.

Acquisition for association rules of information systems is an important problem of knowledge treatment and data mining, and rough sets theory is an important mathematical tools of rules mining. The paper researches the algorithm for optimal minimal mining rules in information systems. By use of the characteristic of elements in association matrix, the algorithm proposed in this paper can find the conditional elements of potential association rules including in the information systems after transforming the property elements of association matrix. The algorithm can directly simplify the mining progress of optimal minimal rules of information systems, and effectively void the NP-hard problem often occurring in the attributes reduction of information systems. An example is also given to prove the efficient lastly.