基于FP-Tree的最大频繁项目集更新挖掘算法

An algorithm of the updating mining for mining maximum frequent itemsets based on FP-Tree

发现最大频繁项目集是多种数据挖掘应用中的重要问题 .在应用中用户需要调整最小支持度 ,以发现更有用的最大频繁项目集 .为此提出了一种最大频繁项目集更新算法 (UMFPA) ,该算法通过对频繁模式树(FP Tree)中的频繁项目头表 (HTable)增加两个域 ,从而将减少在数据库不变而最小支持度变化的情况下的更新挖掘最大频繁项目集的费用 .实验结果表明 。

To search more useful maximum frequent itemsets, the minimum support was adjusted based on the needs in the applications. An efficient updating algorithm UMFPA for mining maximum frequent itemsets was presented. The algorithm improved frequent pattern tree by adding two new fields in the frequent-item header table (HTable) of the FP-Tree, so that it made use of previous mining result to cut down the cost of finding new maximum frequent itemsets based on changeable minimum support in an unchangeable database. An example was given to demonstrate and analyzed the algorithm.