一种优化的关联规则增量更新算法

AN OPTIMIZED INCREMENTAL UPDATING ALGORITHM FOR ASSOCIATION RULES

针对PFUP算法存在扫描多次数据库这个瓶颈问题,提出一种优化的关联规则增量更新算法MIFUP(Mixed Improve Fast Updating)。该算法提出了两种优化策略:借鉴事务压缩原理和用数组存放一阶非频繁项集个数。实验仿真说明,MIFUP算法效率明显优于PFUP算法。

To solve the bottleneck problem of PFUP algorithm which scans database many times, a new improved optimized incremental up- dating algorithm for association rules, MIFUP, is presented. In the executing process of MIFUP algorithm, it adopts two optimized strategies. One is the transaction compress theory, and the other is the array for storing the number of first-order non-frequent item sets. Finally, the simulation experiment shows that MIFUP algorithm is more efficient than PFUP algorithm.