摘要
针对以往的频繁项集增量式更新算法需要多次扫描原数据集,并产生大量冗余的候选项集,本文提出了一种快速的增量更新算法Index-FUP。该算法采用改进后的索引数组挖掘算法得到频繁项集,减少了候选集产生的数量和扫描原数据库次数。理论分析与实验结果证明,在事务数据库和最小支持度同时变化时,该算法具有较高更新效率和灵活性。
Since the existing incremental updating algorithms of frequent itemsets require to scan the database many times, producing many redundant candidate items, we present a fast incremental updating algorithm, named the Index-FUP, to handle these two problems. This algorithm obtains the frequent itemsets by adopting an improved index array data mining algorithm, which can reduce the number of candidate itemsets and times of the databases scanning. Theoretical analysis and experimental results show that the algorithm is efficient and feasible when the minimum support change simultaneously with the transaction database.
出处
《微计算机信息》
2010年第33期150-151,172,共3页
Control & Automation
关键词
索引数组
关联规则
频繁项集
Index array
association rule
frequent itemset
