基于有序FP树和二维列表的频繁模式挖掘算法

Study on frequent patterns mining based on sorted FP-Tree and two-dimensional table

关联法是数据挖掘算法中一种重要的技术,FP-Growth算法是当前最有效的关联法则挖掘算法,主要针对传统的FP-Growth算法当前的一些不足进行改进,提出了一种新的挖掘算法OFP树挖掘算法.一是采用了有序FP树代替传统的FP树,减少存储空间的使用,二是采用二维列表记录项的频繁度,省去为寻找第一次条件模式基而遍历FP树的过程.实验结果表明该算法优于传统FPGrowth算法.

Association rules algorithm is an important technology in data mining algorithms.And FP- Growth algorithm is the most effective association rule mining algorithm. This pa-per presented an algorithm, OFP mine, which mainly aims at the improvement of the tradition-al FP - Growth algorithm . One was using a sorted FP tree instead of a traditional FP tree toreduce the practicality of storage space. Second was using two-dimensional table entries torecord items frequently, which for saving the process to search for the first conditional patternbase by traversing the FP tree. Finally, the experimental evaluation showed that the algorithmwas superior to FP - Growth algorithm.