基于频繁模式树的一种关联规则挖掘算法及其在铁路隧道安全管理中的应用

关联规则的FP-growth算法是数据挖掘中性能较好的一种算法,笔者在分析该算法的基础上进行改造探讨,并提出了一种基于FP-tree的高性能关联规则挖掘算法FP-growthN,该新算法特别适合对那些数据量很大但数据项很稀疏的数据进行挖掘。将新算法用于挖掘铁路隧道各病害的关联中,通过对成都铁路局管辖的2005年的2787条隧道病害数据的343条重点隧道有效病害数据的关联分析,得出了各隧道病害之间隐藏着的关系。新法的提出及其应用结果对铁路部门制定检测标准和防治隧道病害有一定的指导作用。

A High-performance Association Rule Mining Algorithm Based on FP-tree and Its Application in Railway Tunnel Safety Management

FP-growth (Frequent Patterns-growth) algorithm for association rules is the one with relatively good performance in data mining. On the basis of the analysis and the innovation of this algorithm, a new and high-performance algorithm of FP-growthN was built according to FP-tree (Frequent Patterns-tree), which was especially suitable for mining these data with large volume yet sparse items. Then, this new algorithm was applied to mining the association of damages in railway tunnels. The hidden relations among tunnel damage were discovered through the association analysis of valid damage data from 343 out of 2787 tunnels damaged data in 2005, which are ruled over by the Chengdu Railway Bureau. The result obtained by the new algorithm is helpful for the prevention of damages and the establishment of detection criterion in tunnel industry.