为了解决Apriori算法中生成候选二项集不能剪枝、多次扫描数据库中对计算支持度不产生作用的项等缺点,提出了使用矩阵直接生成频繁二项集及候选三项集,使用链表删除对计算支持度不产生作用的数据的方法。通过设置不同的最小支持度将两个算法的运行时间进行对比,根据结果分析证明改进后的算法比Apriori算法执行效率高。%In order to solve these disadvantages,like generating candidate binomial spaces can't be pruned,and scanning the terms that have no effect on the counting support in database for many times,then,proposing methods of using Matrix to generate Frequent binomial spaces and candidate trinomial spaces directly,and using Linked List to delete the terms that have no effect on the counting support.Last but not the least,this paper validates the improved algorithm by setting different minmum support of two algorithms to compare the running time to prove that the execution efficiency of the algorithm can be increased.
展开▼