摘要
该文针对K均值聚类算法存在的缺点,提出一种改进的粒子群优化(PSO)和K均值混合聚类算法。该算法在运行过程中通过引入小概率随机变异操作增强种群的多样性,提高了混合聚类算法全局搜索能力,并根据群体适应度方差来确定K均值算法操作时机,增强算法局部精确搜索能力的同时缩短了收敛时间。将此算法与K均值聚类算法、基于PSO聚类算法和基于传统的粒子群K均值聚类算法进行比较,数据实验证明,该算法有较好的全局收敛性,不仅能有效地克服其他算法易陷入局部极小值的缺点,而且全局收敛能力和收敛速度都有显著提高。
To deal with the problem of premature convergence of the traditional K-means algorithm, a novel K-means cluster method based on the enhanced Particle Swarm Optimization（PSO） algorithm is presented. In this approach, the stochastic mutation operation is introduced into the PSO evolution, which reinforces the exploitation of global optimum of the PSO algorithm. In order to avoid the premature convergence and speed up the convergence, traditional K-means algorithm is used to explore the local search space more efficiently dynamically according to the variation of the particle swarm＇s fitness variance. Comparison of the performance of the proposed approach with the cluster method based on K-means, traditional PSO algorithm and other PSO-K-means Mgorithm is experimented. The experimental results show the proposed method can not only effectively solve the premature convergence problem, but also significantly speed up the convergence.
出处
《电子与信息学报》
EI
CSCD
北大核心
2010年第1期92-97,共6页
Journal of Electronics & Information Technology
基金
哈尔滨工程大学校科研基金(002080260735)
黑龙江省博士后基金(LBH-Z08227)资助课题
关键词
K均值算法
粒子群优化算法
随机变异
适应度方差
K-means algorithm
Particle Swarm Optimization（PSO） algorithm
Stochastic mutation
Fitness variance
