一个新的最优简单规则挖掘算法  

New Algorithm for Optimal Minimal Rules Mining

作　　者：邱卫根[1,2] 徐相林[1,3] 

机构地区：[1]广东工业大学计算机学院,广东广州510090 [2]智能技术与系统国家重点实验室,计算机科学与技术系,北京100084 [3]广州工程技术职业学院,广东广州510090

出　　处：《小型微型计算机系统》2009年第3期417-420,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60474072)资助;广东省自然科学基金项目(04009465)资助

摘　　要：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.

关 键 词：最简规则 粗集理论 关联矩阵 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]