基于快速模式增长的FP-growth改进算法  

Modified FP-growth Algorithm Based on Rapid Pattern Growth

作　　者：党敏侠[1] 孟德鑫[1] 

机构地区：[1]中国电子科技集团公司第二十八研究所,南京210007

出　　处：《指挥信息系统与技术》2012年第4期73-76,共4页Command Information System and Technology

摘　　要：FP-growth算法是当前一种高效的频繁模式挖掘算法,分析了FP-growth算法的性能消耗,引入了条件频度基的数据结构用于存储构建条件FP-tree的所有频度信息,提出了自顶向下与自底向上结合的模式增长方法,该方法避免了原算法在遍历FP-tree时的性能浪费。试验表明,运用了快速模式增长的改进算法的效率比原算法显著提高。The FP-growth algorithm is currently an efficient algorithm for frequent pattern mining.The performance consumption of the FP-growth algorithm is analyzed.A data structure of conditional frequency base is introduced to store all frequency information for building condition FP-tree.The combined up-bottom and bottom-up pattern growth method is proposed to avoid the performance waste when traversing the FP-tree.Experimental results show that the efficiency of the rapid pattern growth is improved than that of the original algorithm.

关 键 词：频繁模式挖掘 FP-GROWTH算法 条件频度基 模式增长 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]