网格环境下基于WEKA4WS的分布式矩阵关联规则挖掘算法

WEKA4WS-Based Distributed Mining Algorithm for Matrix Association Rules Under Grid Environment

在相关矩阵理论的研究基础上,将网格与Web服务技术融合,以分布式问题求解环境和开源数据挖掘类库weka为底层支撑,构建了网格环境下面向服务的分布式数据挖掘体系,提出一种基于矩阵的分布式关联规则算法.该算法不需要进行复杂的寻找频繁项集的过程,直接通过关联矩阵就可以判断出,给出了算法的理论证明,并通过实例验证了算法的正确性、有效性和体系结构的可行性,对于解决分布式关联规则挖掘问题有了一个新的突破.

On the basis of in-depth research of the matrix theory and the association of grid and Web service technology,the paper presents a service-oriented distributed data mining system under grid environment,which is supported by distributed solution environment and Weka library.A matrix-based association rule algorithm is given,which can be judged directly by association matrix rather than by complicated seeking of frequent item sets.The algorithm is theoretically and practically proved to be correct,valid and feasible,which indicates a breakthrough in dealing with the distributed mining problem of association rule.