摘要
K-Means算法是一种经典的基于划分的聚类方法。传统的K-Means算法中存在很明显的缺陷,它对初始聚类中心的依赖性很大,聚类结果很容易陷入局部最优值;而基于遗传算法改进的K-Means聚类方法,提高了聚类结果的稳定性,但因为个体的多样性不足,常常会出现早熟等现象,其局部寻优能力较弱。针对上述问题,文中提出一种结合模拟退火算法的遗传K-Means聚类方法。利用模拟退火算法改进遗传算法的变异操作,用K-Means操作取代遗传算法的交叉操作,改善早熟现象,避免聚类结果陷入局部最优,实现聚类方法性能的提升。实验结果表明,该方法的聚类准确度比一般K-Means方法和遗传K-Means方法都要高。
K-Means algorithm is one of the most classical division-based clustering methods. In the traditional K-Means algorithm,there are obvious flaws like strong dependence on the initial clustering center and the clustering result is easy to fall into the local optimal value. The improved K-Means clustering method based on genetic algorithm improves the stability of clustering results. However,due to the insufficient diversity of individuals,prematurity and other phenomena often occur,and its local optimization is weak. For this,we present a genetic K-Means clustering method combined with simulated annealing algorithm. The simulated annealing algorithm is used to improve the mutation operation of genetic algorithm,the classical K-Means operation is used to replace the crossover operation of the genetic algorithm,so as to improve the premature phenomenon,avoid the clustering result falling into the local optimal,and improve the performance of the clustering method. The experiment shows that the clustering accuracy of the proposed method is higher than that of the general K-Means method and the genetic K-Means method.
作者
凌静
江凌云
赵迎
LING Jing;JIANG Ling-yun;ZHAO Ying(School of Telecommunications & Information Engineering,Nanjing University of Posts and Telecommunications,Nanjing 210003,China)
出处
《计算机技术与发展》
2019年第9期61-65,共5页
Computer Technology and Development
基金
国家自然科学基金(6127123)
关键词
聚类
K-MEANS算法
遗传算法
模拟退火算法
clustering
K-Means algorithm
genetic algorithm
simulated annealing algorithm