网格环境下基于Weka4WS的分布式聚类算法*

Weka4WS采用WSRF技术用于执行远程的数据挖掘和管理分布式计算，支持分布式数据挖掘任务。基于Weka4WS和网格环境，尝试了一种新的分布式聚类方法,并成功地将其嵌入到Weka4WS框架中,借助Weka Library实现分布式数据挖掘算法,同时引入了距离代价和混合概率的概念,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为底层支持环境,构建了网格环境下面向服务的分布式数据挖掘体系,并以基于Weka4WS的分布式聚类算法验证了算法的有效性和体系结构的可行性。

Research on distributed clustering algorithm Weka4WS-based in grid environment

A grey particle swarm optimization algorithm based on TOPSIS for solving multi-objective optimization problems, which proposed by taking advantage of technique for order preference by similarity to ideal solution (TOPSIS) trace Pareto front for distance and grey correlation degree distinguish similarity between curves of non-inferior solution sets and curves of Pareto front solution sets. The algorithm took advantage of TOPSIS theory and grey correlation degree theory to acquire relative fitness coefficient and gray correlation coefficient, and defined their sum as relatively ideal degree, which distinguished advantages and disadvantages of particles and determines individual extreme and global extreme. Validated the algorithm using four different types benchmark cases. The experimental results show that grey particle swarm optimization based on TOPSIS, compared with objective weighting method and grey PSO algorithms, it can find many Pareto optimal solutions distributed onto the Pareto front and do not increase the complexity of the algorithm.