分布式动态数据库增量关联规则挖掘研究

Research on Incremental Association Rules Mining in Distributed Dynamic Database

为了解决分布式动态数据库关联规则挖掘效率低的问题,利用MPI与OpenMP的优点,提出了实现增量关联规则挖掘的混合模式。在次频繁项概念基础上,给出该混合模式总体架构,设计了基于MPI与OpenMP的分布式动态数据库增量关联规则挖掘混合模式工作流程,并给出了伪代码描述,该模式只处理变化的数据。实验结果表明,该模式比现有的串行与分布式关联规则挖掘方法效率更高、性能更优。

In order to solve the problem of low efficiency of association rules mining methods in the existing distributed dynamic database, a hybrid model of incremental association rule mining approach is provided based on the advantages of MPI and OpenMP in this paper. The overall architecture of the hybrid model is proposed based on pre large concept, the workflow of incremental association rules mining based on MPI and OpenMP is designed, and the pseudo code description is given. The model is to process only new data changes. Experimental results show that the hybrid approach is more efficient and more performance than the existing serial and distributed association rules mining methods.