基于改进GWO-CV优化的K-调和均值聚类算法  

K-harmonic Mean Clustering Algorithm Based on Improved GWO-CV Optimization

作　　者：张文宇 张茜 杨媛 刘嘉 Zhang Wenyu;Zhang Xi;Yang Yuan;Liu Jia(College of Economics and Management,Xi'an University of Posts&Telecommunications,Xi'an 710061,China;China Research Institute of Aerospace Systems Science and Engineering,Beijing 100854,China)

机构地区：[1]西安邮电大学经济与管理学院,西安710061 [2]中国航天系统科学与工程研究院,北京100854

出　　处：《统计与决策》2020年第16期9-13,共5页Statistics & Decision

基　　金：陕西省教育厅重点项目(19JZ056)。

摘　　要：为克服传统聚类算法对初始聚类中心敏感且容易陷入局部最优的问题,文章提出一种基于改进的灰狼优化与交叉验证法结合的K-调和均值聚类算法(GWO-CVKHM)。首先将新的非线性收敛因子引入灰狼优化算法,以调整前期广度搜索与后期深度搜索比例,同时基于模糊控制权重决策对灰狼种群位置进行更新;其次利用改进灰狼优化算法与交叉验证的思想对初始聚类中心进行寻优;最后基于改进后的聚类算法选取UCI数据库中真实数据集进行聚类。实验结果表明,该算法在求解精度及算法稳定性方面优于对比算法,具有更快的收敛速度与更强的全局搜索能力。In order to solve the problem that the traditional clustering algorithm is sensitive to the initial clustering center and easy to fall into local optimum, this paper proposes a k-harmonic mean clustering algorithm(GWO-CVKHM) based on the combination of improved gray wolf optimization and cross validation method. Firstly, the new nonlinear convergence factor is introduced into the gray wolf optimization algorithm to adjust the proportion of early extensive search and late deep search, and at the same time, the position of gray wolf population is updated based on fuzzy control weight decision-making. Secondly, the initial clustering center is optimized by using the improved gray wolf optimization algorithm and cross validation. Finally, based on the improved clustering algorithm, the real data sets in UCI database are selected to be clustered. Experimental results show that the proposed algorithm is better than the comparative algorithm in terms of solution accuracy and stability, and has faster convergence speed and stronger global search ability.

关 键 词：K-调和均值聚类 灰狼优化算法(GWO) 交叉验证法(CV) 全局搜索能力 

分 类 号：O21[理学—概率论与数理统计]