基于多种群协同进化算法的数据并行聚类算法  

Data parallel clustering algorithm based on multi-group co-evolution algorithm

作　　者：孙柳 SUN Liu(Huali College Guangdong University of Technology,Guangzhou 511325,China)

机构地区：[1]广东工业大学华立学院,广州511325

出　　处：《智能计算机与应用》2021年第6期144-147,152,共5页Intelligent Computer and Applications

摘　　要：为了提高云存储空间多维资源数据挖掘能力,需要进行数据并行聚类处理,提出基于多种群协同进化算法的数据并行聚类算法。构建云存储空间多维资源数据的参数采集模型,对采集的云存储空间多维资源数据进行模糊并行特征分布式重组,提取云存储空间多维资源数据聚类特征参数集,采用关联粗糙集特征分析方法进行云存储空间多维资源数据的多尺度小波结构分解,结合多种群协同控制的方法,建立云存储空间多维资源数据的并行聚类模型,通过关联协同滤波检测方法,进行云存储空间多维资源数据的分组特征检测和融合聚类处理,根据差分进化方法进行云存储空间多维资源数据的聚类中心寻优,遍历云存储空间多维资源数据聚类区域的候选目标集实现对云存储空间多维资源数据的并行关联规则聚类和可靠性挖掘。仿真结果表明,采用该方法进行云存储空间多维资源数据挖掘的精度较高聚类性较好,收敛性较强。In order to improve the data mining ability of multi-dimensional resources in cloud storage space,data parallel clustering processing is needed,and a data parallel clustering algorithm based on multi-group co-evolution algorithm is proposed.Construct a parameter collection model of cloud storage space multi-dimensional resource data,perform fuzzy parallel feature distributed reorganization of the collected cloud storage space multi-dimensional resource data,extract cloud storage space multi-dimensional resource data clustering feature parameter set,and use the associated rough set feature analysis method Multi-scale wavelet structure decomposition of cloud storage space multi-dimensional resource data,combined with multiple group collaborative control methods,establishes a parallel clustering model of cloud storage space multi-dimensional resource data,and grouping cloud storage space multi-dimensional resource data through associated collaborative filtering detection methods Feature detection and fusion clustering processing,according to the differential evolution method to optimize the clustering center of cloud storage space multi-dimensional resource data,traverse the candidate target set of the cloud storage space multi-dimensional resource data clustering area,and realize the cloud storage space multi-dimensional resource data Parallel association rule clustering and reliability mining.The simulation results show that this method has higher precision,better clustering and stronger convergence for multi-dimensional resource data mining in cloud storage space.

关 键 词：多种群 协同进化 云存储 数据 并行聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]