基于聚类分析的改进微粒群算法  

Improved Particle Swarm Optimization Based On Cluster Analysis

作　　者：田雅娟[1] 贺文彬[1] 赵子强[1] 

机构地区：[1]山西中医学院,山西太原030024

出　　处：《山西师范大学学报（自然科学版）》2013年第4期27-31,共5页Journal of Shanxi Normal University(Natural Science Edition)

摘　　要：聚类分析是依据样本间关联的量度标准将其自动分成几个群组,使同一群组内的样本相似,而属于不同群组的样本相异的一种方法.在微粒群算法中由数量不等的粒子根据规则组合成不同的群体,所有的群体最终将会向着一个全局最优的位置运动.本文将通过改进微粒群算法的局部更新规则来改善算法的性能,根据由聚类半径确定初始聚类中心的方法将粒子群进行分类,然后运用该方法对所有粒子进行分类,初始化得到不同的粒子群体,最后对整个粒子群体进行优化得到全局最优解.Clustering analysis is based on a measure of the correlation between sample standards, which will be automatically divided into several groups, and the samples in the same group are similar, but other samples of different groups are dissimilar. In particle swarm optimization, the number of particles combined into different groups according to the rules, all groups will eventually towards a global optimal position movement. In this arti- cle, we make local update rules of the improved particle swarm algorithm to improve the performance of the algo- rithm, determine the initial clustering center according to the clustering radius, and categorize the whole particle swarm, then we use the method to classify all the particles, initialize the particles of different groups, finally op- timize the whole particle groups to get the global optimal solution.

关 键 词：数据挖掘 聚类分析 微粒群算法 

分 类 号：O152[理学—数学]