In the information and data explosion era,redundant rules problem has become a major obstacle to gain knowledge for data miners,but the current solution may lead to the incompleteness of association rules. For the issues above,introduce the directed hyperg-raph to express association rules,redefine the adjacency matrix,propose the categorical thinking for removing redundant rules,dividing re-dundant rules into slave rules and repeated-path-rules,the algorithm of removing the dependency rule by VB programming and the span-ning tree algorithm to remove the redundancy of repeated-path-rules. The result of experiments show that this method is effective,fast and to maintain the integrity and accuracy of association rules which links directed hypergraph,spanning tree of graph theory and associa-tion rules innovatively. At the same time it removes all redundant rules.%在信息及数据爆炸的时代，冗余问题已经成为数据挖掘者获得知识的重大障碍，而目前解决的方法会导致关联规则的不完整性。基于此，文中引入了有向超图表示关联规则，重定义了邻接矩阵，介绍了冗余规则分类处理思想，将冗余规则分为从属规则和重复路径规则，通过VB编程去除了从属规则冗余，以及利用生成树算法去除了重复路径规则冗余。实验结果证明，此方法创新性地结合了图论中有向超图、生成树与关联规则的知识，维护了关联规则的完整性和准确性，同时去除了全部冗余规则。
展开▼
