频繁项集挖掘是数据挖掘的一个热点,频繁项集的精简表示有助于减少频繁项集的数量,是频繁项集挖掘中的一个关键性问题。文章在总结频繁项集的相关精简模型后,对频繁项集精简模型,生成器精简表示模型进行研究。传统的生成器精简表示模型每次生成都需要扫描数据库,效率不高。本文提出一种新颖的频繁项集精简算法 FPASCAL,使用 FP 树数据结构代替数据库再结合剪枝策略。理论分析及实验证明该方法在时间性能和空间复杂度上都优于传统方法。%Frequent itemset mining is a very hot topic in data mining.The concise representation of frequent itemset is a key problem, as it can reduce the frequent itemset numbers without information less .After summarizing related frequent itemset concise representation models ,we research a frequent itemset concise model and its generator representation .While the traditional genera-tor representation need scanning database which makes it low efficient , a new frequent itemset concise representation FPASCAL al-gorithm is put forwarded ,after using FP tree instead of data base and combine some pruning strategy .The experiment result proves that the FPASCAL algorithm is more better than the traditional algorithm both time consuming and space consuming .
展开▼
