基于FP-growth的数据关联改进算法

Advanced data association algorithm based on FP-growth

随着现如今数据收集能力和存储能力的大大增强,大规模数据挖掘分析的重要性越来越显得重要。然而,对大规模数据的分析挖掘并不是一件容易的事情。因此,为了可以更高效的分析这些数据,很多新的算法和数据结构逐渐被引入到了数据挖掘分析中去。针对关联分析,提出了一种名为高效频繁模式挖掘(advanced frequent pattern mining,AFPM)算法。基于前置频繁模式树(pre-frequent pattern tree,PFP-tree)来提升关联分析的性能,并提供了相应的算法来实现基于这种数据结构的关联分析。通过大量的实验数据验证了这种新型的数据结构在关联分析问题上是优于频繁模式增长(FP-growth)算法。

With the current data collection capacity and storage capacity greatly enhanced, the importance of large-scale data mining analysis more and more important. However, the analysis of large-scale data mining is not an easy thing. Therefore, in order to be able to more efficient analysis of these data, many new algorithms and data structures are gradually introduced to the data mining analysis. This paper is based on the correlation analysis, based on this article, proposed a called advanced frequent pattern mining （AFPM） algorithm. This algorithm is based on the pre-frequent pattern tree （PFP-tree） to improve the performance of association analysis and provide the corresponding algorithm to implement the association analysis based on this data structure. It is proved that this new data structure is superior to FP-growth algorithm in association analysis problem through a large number of experimental data.