计算机工程与应用 ›› 2019, Vol. 55 ›› Issue (18): 140-145.DOI: 10.3778/j.issn.1002-8331.1904-0280

• 模式识别与人工智能 • 上一篇    下一篇

基于改进粒子群算法的[k]均值聚类算法

汤深伟,贾瑞玉   

  1. 安徽大学 计算机科学与技术学院,合肥 230601
  • 出版日期:2019-09-15 发布日期:2019-09-11

[K]-Means Cluster Algorithm Based on Improved PSO

TANG Shenwei, JIA Ruiyu   

  1. School of Computer Science and Technology, Anhui University, Hefei 230601, China
  • Online:2019-09-15 Published:2019-09-11

摘要: 针对[k]-means算法易受初始中心影响的缺点,提出了基于改进粒子群算法的[k]-means聚类算法[(k]-means cluster algorithm based on Improved PSO,IPK-means),在粒子群算法中加入混沌搜索过程,以增加PSO迭代后期粒子群的多样性,并且在粒子更新过程中,给出了一种动态调整因子公式,使得调整因子与该粒子的适应度值大小相关,即同一迭代中不同粒子也会拥有不同的调整因子。最后将改进的PSO算法应用于[k]-means聚类,为其寻找较好的初始中心,实验结果表明了该算法可取得较好的聚类结果。

关键词: 粒子群优化, [k]均值聚类, 混沌搜索, 自适应调整因子

Abstract: The [k]-means cluster algorithm based on Improved PSO(IPK-means) is proposed for [k]-means algorithm’s disadvantage that it is vulnerable to the influence of the initial center, adding chaotic search process to the Particle Swarm Optimization(PSO) algorithm in order to increase the PSO iteration late particle swarm diversity, and in the process of particle update, it proposes a dynamic adjustment factor formula, which makes the adjustment factor related to the fitness value of the particle size, different particles in the same iteration also have different adjustment factors. Finally, the improved PSO algorithm is applied to [k]-means clustering to find a better initial center for it. The experimental results show that this algorithm can achieve better clustering results.

Key words: particle swarm optimization, [k]-means cluster, chaotic searching, adaptive adjustment factor