一种k-means聚类的改进算法与实现  

An Improved K-Means Clustering Algorithm and Its Implementation

作　　者：冯能山[1] 林志华[1] 熊金志[1] 祝建军[2] 

机构地区：[1]东莞理工学院计算机学院,广东东莞523808 [2]华南理工大学计算机科学与工程学院,510641

出　　处：《软件导刊》2012年第3期66-70,共5页Software Guide

基　　金：广东省自然科学基金项目(9151170003000017)

摘　　要：传统的k-means算法作为一种动态聚类法,是聚类方法中常用的一种划分方法,其应用领域非常广泛。但该方法存在初始k值不确定、时间复杂度大等缺点。针对这些缺点,改进了聚类初值的随机性问题,简化了算法,降低了时间复杂度,提高了k-means算法的性能,并给出了具体的代码实现。As a dynamic clustering method,the traditional algorithms of k-means is a common way of division in clustering with its wide applications in a range of fields.However,this means is flawed for a series of shortcomings with it including the uncertain initial value in k and the high degree of time complexity.In order to overcome these problems,this paper contributes to make improvements in the resolution of the random problem of initial value for clustering and the simplification for the algorithm as well as the reduction in the degree of time complexity.Furthermore,the paper also enhances the performance of the algorithms of k-means and provides a concrete code for implementation.

关 键 词：动态聚类 K-MEANS算法 初值 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]