一种基于单处理机的并行关联规则算法及其在数字图书馆中的应用

One Parallel Association Rule Algorithm Based on Uniprocessor and its Application in Digital Library

针对现有并行挖掘技术需要建立在专门的并行机上而使中小图书馆难以实现的弊端,通过研究关联规则的内在并行性,提出一种基于图形处理单元技术的快速并行关联规则算法,实现在单处理器上并行数据挖掘。通过仿真实验证明,优化后的算法在不增加硬件设备的前提下实现单处理机并行,提高运行速度,解决传统并行关联规则挖掘难以应用化的问题,该算法对于数字图书馆具有重要的理论和应用价值。

In view of the existing parallel mining technology＇ s requirement of establishing on special parallel machine and the defect of medium and small-scale library hard to realize it, a rapid parallel association rule algorithm is proposed based on graphic processing unit technology, which realizes parallel mining on uniprocessor. Through the proof of simulation experiment, on the premise of not increasing hardware equipment, the optimized algorithm realizes uniprocessor parallel and raises the speed, solving the problem of traditional parallel association rule hard to engineering. This algorithm has important theoretical significance and application value for digital library.