一种基于FP-树的时态关联规则的分区挖掘方法  

A FP-tree Based Partition Mining Approach to Discovering Temporal Association Rules

作　　者：马慧[1] 汤庸[1] 潘炎[1] 

机构地区：[1]中山大学计算机科学系,广州510275

出　　处：《计算机工程》2006年第17期132-134,共3页Computer Engineering

基　　金：国家自然科学基金资助项目(60373081);广东省自然科学基金资助重点项目(04105503)

摘　　要：随着各种形式的数据的迅速增长,业务数据中的时态信息挖掘问题受到人们普遍关注。该文提出了一种带有效时间区间的时态关联规则,给出了一种基于FP-树的挖掘方法。该方法利用分区挖掘的思想,以分区为单位表示项集的有效时间区间,并为每个分区构建FP-树,大大简化了对某个项集在其有效时间区间中的出现次数的计算,从而更有效地计算时态置信度。最后用一个例子对该方法的执行过程进行了阐述。With the rapid growth of data available from all kinds of sources, temporal information mining in business data has been a hot area attracting more and more attention. An approach to discover temporal association rules within valid time intervals is investigated. An algorithm based on FP-tree is devised. In this approach, by means of partition mining, each itemset is associated with a valid time interval presented by units of partitions. For each partition, a FP-tree is constructed to help the calculation of the count of an itemset within its valid time interval, which can ,smooth the calculation of temporal confidence, At last, an example is given to demonstrate the mining process.

关 键 词：数据挖掘 时念数据挖掘 关联规则 频繁模式树 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]