动态遗传算法在大数据聚类分析中的应用  

A Dynamic Clustering Method Based on Genetic Algorithms For BIG Data Analysis

作　　者：向丽 戴晓晖[2] Xiang Li;Dai Xiaohui(Shenzhen Konka Information Network Co.,td Shenzhen,Guangdong 518053,China;Petro-CyberWorks Information Technology Co.,Ltd.Beijing 100086,China)

机构地区：[1]深圳康佳信息网络有限公司,广东深圳518053 [2]石化盈科信息技术有限责任公司,北京100086

出　　处：《信息记录材料》2019年第8期19-21,共3页Information Recording Materials

摘　　要：对于常规聚类方法,聚类结果往往与初始聚类中心数目和数据入次序有关。本文另辟蹊径,提出了一种动态GA来实现样本类别数目由数据本身来确定,避免了聚类数目确定的盲目性等问题。计算结果表明,该方法是一个具有全局最优解的动态聚类方法。To solve the problem of sensitivity with original clustering center;clustering results depended on the order of inputted example in common dynamic clustering algorithm,and blindness of decided the number of clusters,a new dynamic clustering method based on genetic algorithms is presented in this paper and is applied for data auto clustering in databases.Computed results indicate that the method is a dynamic clustering algorithm with global optimization.

关 键 词：遗传算法 聚类分析 多维空间 大数据 全局优化 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]