基于最大距离积与最小距离和协同K聚类算法

摘要

An improved K-clustering algorithm based on the maximum distance product and the sum of the minimum distance was proposed,which solved the problems that the traditional K-means algorithm had a large randomness, poor stability and a maximum distance product method with a large number of iterations and a long operation time problem. Firstly,the algorithm calculated the density parameters through the distribution of the samples, and then built a high-density point set.Based on this, a high-density object farthest from the sample center was regarded as the first initial cluster center,and the rest of the initial cluster centers were obtained by the maximum distance product method.In the process of cluster center update,the data objects with the smallest distance to the sample in the cluster were selected as the cluster centers,and then the other data objects were divided into the corresponding clusters by the minimum distance so as to realize the clustering.The experimental results on the UCI dataset showed that the proposed cooperative K-clustering algorithm had faster convergence rate,more accurate clustering result and higher stability than the other two improved algorithms and the K-means algorithm.%提出一种基于最大距离积与最小距离之和的协同K 聚类改进算法,解决了传统K-means 算法聚类结果随机性大?稳定性差,以及最大距离乘积法迭代次数多?运算耗时长等问题?该算法首先通过样本的分布情况计算其密度参数,进而构建高密度点集合,在此基础上将距离样本集中心最远的高密度对象作为第一个初始聚类中心,再通过最大距离乘积法求得其余初始聚类中心;在簇中心更新过程中,选取与簇内样本距离之和最小的数据对象作为簇中心,再将其他数据对象按最小距离划分到相应簇中, 从而实现聚类?在UCI 数据集上的实验结果表明,与其他两种改进算法以及K-means 算法相比,新提出的协同K 聚类算法具有更快的收敛速度?更准确的聚类结果和更高的稳定性.