基于布谷鸟算法的K-medoids 聚类挖掘与并行优化  

K-medoids Clustering Mining and Parallel Optimization Based on Cuckoo Algorithm

作　　者：谭成兵[1] 刘源 徐健 Tan Chengbing;Liu Yuan;Xu Jian(Information Technology Department,Bozhou Vocational and Technical College,Bozhou 236813,China;Information Center,Guilin Medical University,Guilin 541001,China;School of Mechanical and Electrical Engineering,Guilin University of Electronic Technology,Guilin 541001,China)

机构地区：[1]亳州职业技术学院智能工程系,安徽亳州236813 [2]桂林医学院信息中心,广西桂林541001 [3]桂林电子科技大学机电工程学院,广西桂林541004

出　　处：《台州学院学报》2021年第3期7-12,共6页Journal of Taizhou University

基　　金：安徽省高校自然科学研究重点资助项目(KJ2018A0881);亳州市科技重大专项;桂林医学院基金项目(31304019011)。

摘　　要：为了提高聚类挖掘的准确度和效率,采用布谷鸟优化的K-medoids算法来完成聚类,通过多节点并行聚类提高聚类效率。随机选择K-medoids聚类中心,并构建适应度函数,根据聚类样本初始化布谷鸟鸟巢位置。在布谷鸟K-medoids聚类中,可采用多运算节点并行聚类方法获取最优解,以优化聚类时间。实验证明,通过合理设置宿主发现概率阈值,布谷鸟优化的K-medoids聚类在自有数据集和UCI公开数据集聚类中均表现出良好的聚类准确率,比常用聚类算法的聚类准确率更高,采用多节点并行优化后,提高了大规模数据样本的聚类效率。To improve the performance of clustering mining,cuckoo-optimized K-medoids algorithm is proposed to complete clustering,and multi-node parallel clustering is adopted to improve clustering efficiency.The K-medoids clustering center is randomly selected and the fitness function is established.Cuckoo nest position is initialized based on clustering samples.In the process of clustering,the parallel clustering method with multiple computing nodes could be used to efficiently obtain the optimal solution.Experiments represent that the cuckoo-optimized K-medoids clustering shows good clustering accuracy in both self-owned data sets and UCI public data sets by setting the threshold of host discovery probability reasonably.

关 键 词：聚类 K-medoids算法 布谷鸟算法 并行优化 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] TP18[自动化与计算机技术—计算机科学与技术]