摘要
Han等人提出了频繁模式增长FP-growth算法,该算法在第一次扫描数据库后,得到频繁项集合和每个频繁项的支持度,并按支持度降序排列,但没有对支持度相同项的排列做进一步说明。本文依据"越是频繁出现,越可能被共享"的建树原则,提出了通过比较该项与前后项项集的支持度,较大者先排列的方法,使后续构建的FP-tree比任意排序构建的FP-tree更优。
Han et al. propose the FP - growth algorithm of Frequent pattern growth, after which scans the database for the first time , obtained the frequent item sets and support of each frequent item, and arranged in descending order ,but don' t illustrate for further for the arrangement of the same support item. This article proposes a method that the first item arranged by comparing the item set support of this item with the before and after, according to the principle of "the more frequent, the more likely to be shared" , so that the latter FP - tree by this method is better than by random arrangement.
出处
《网络新媒体技术》
2012年第4期53-56,共4页
Network New Media Technology