传统与流数据聚类算法  

Traditional Data Stream Clustering Algorithm

作　　者：刘晓璐 王志栋 单广荣[1] LIU Xiao-lu;WANG Zhi-dong;SHAN Guang-rong(Key Laboratory of China's Ethnic Languages and Information Technology of Ministry of Education,Northwest Minzu University,Lanzhou 730030;School of Vehicle Engineering,Dalian Jiaotong University,Dalian 116028)

机构地区：[1]西北民族大学,甘肃省民族语言智能处理重点实验室,兰州730030 [2]大连交通大学机车车辆工程学院,大连116028

出　　处：《现代计算机》2020年第29期25-28,共4页Modern Computer

摘　　要：在数据骤增的大数据时代,聚类算法成为研究热点。首先,介绍传统聚类算法与流数据聚类算法,流数据聚类算法可以达到数据快速扫描并分类形成簇集合的效果。此外,介绍基于划分的传统算法K-means、K-means++、K-中心点以及基于流数据的Stream算法;基于层次的传统算法BIRCH以及基于流数据的CluStream算法;基于密度的传统算法DNSCAN以及基于流数据的DenStream算法;基于网格的传统算法CLIQUE以及基于流数据的D-Stream算法。In the era of big data with a sudden increase in data,clustering algorithms have become a hot research topic.First,the traditional clustering algorithm and streaming data clustering algorithm are introduced.The streaming data clustering algorithm can achieve the effect of rapid data scanning and classification to form cluster sets.In addition,introduce traditional partition-based algorithms K-means,K-means++,K-medoids algorithm and STREAM algorithm based on streaming data;traditional layer-based algorithm BIRCH algorithm and data stream-based CluStream algorithm;traditional density-based algorithm DNSCAN and DenStream algorithm based on data stream;CLIQUE,a traditional grid-based algorithm,and D-Stream algorithm based on streaming data.

关 键 词：传统聚类 流数据聚类 类 大数据 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]