[1]曲建华,邵增珍.多种群协同进化的K-means聚类算法[J].南京师大学报(自然科学版),2010,33(03):122-126.
　Qu Jianhua,Shao Zengzhen.Cooperative Evolutionary K-means Clustering Algorithm With Multi-populations[J].Journal of Nanjing Normal University(Natural Science Edition),2010,33(03):122-126.
点击复制

多种群协同进化的K-means聚类算法()
分享到：

《南京师大学报（自然科学版）》[ISSN:1001-4616/CN:32-1239/N]

卷:
第33卷
期数:
2010年03期
页码:
122-126
栏目:
计算机科学
出版日期:
2010-09-20

文章信息/Info

Title:
Cooperative Evolutionary K-means Clustering Algorithm With Multi-populations
作者:
曲建华1;  邵增珍2
1. 山东师范大学管理与经济学院, 山东济南250014 2. 山东师范大学信息科学与工程学院, 山东济南250014
Author(s):
Qu Jianhua1; Shao Zengzhen2
1.School of Management and Economics,Shandong Normal University,Jinan 250014,China 2. School of Inform at ion Science and Engin eering, Shandong N orm alUn ivers ity, Jinan 250014, C h ina
关键词:
多种群;  微粒群算法;  K 均值算法;  协同进化
Keywords:
mu lt-i population;  PSO;  K-m eans;  cooperative evo lu tion
分类号:
TP301.6
摘要:
针对K均值聚类算法易陷入局部最小的缺点,提出了一种多种群协同进化的微粒群和K均值混合聚类算法,它将整个种群分解为多个子种群,各子种群独立进化,周期性地更新共享信息.同时将此算法与现有的基于遗传算法的K均值聚类算法进行了比较.实验结果证明,该算法能有效地克服传统的K均值算法易陷入局部极小值的缺点,同时全局收敛能力优于基于遗传算法的K均值聚类算法.
Abstract:
Th is paper presents an m ixed c lustering a lgor ithm based on cooperative evo lution w ith mu lt-i populations. It adopts cooperativ e evo lutionary strategyw ith mu lt-i popu lations to chang e them ode o f trad itional search ing optim um so lutions. The whole cluster ing process is d iv ided into tw o stages. The first stag e uses the cooperative evo lutiona ry PSO algorithm to search the in itia l c lustering centers. The second stage uses theK-m eans algor ithm. The exper im ent prov ed that this m ethod w as ab le to ex trac t the correct number o f c lusters w ith good c lustering qua lity com pa red to the results obta ined from other cluster ing a lgor ithm s likeK-m eans and PSO c luste ring a lgor ithm.

参考文献/References:

[ 1] LiM J, NgM K, Cheung Y M, e t a.l Agg lom era tive fuzzyK-m eans cluster ing algor ithm w ith se lection o f number of c lusters [ J] . IEEE T ransactions on Know ledge and Da ta Eng ineer ing, 2008, 20( 11): 1 519-1 534.
[ 2] LaszloM, M ukher jee S. A gene tic algor ithm that ex changes neighbo ring centers forK-m eans cluster ing [ J]. Pattern Recogn-i tion Lette rs, 2007, 28( 16): 2 359-2 366.
[ 3] ArthurD, Vassilv itskiiS. K-m eans+ + : the advantages of careful seed ing[ C ] / / Proceedings of the 18 th annualACM-S IAM Sym pos ium on D iscrete a lgo rithm s. New Or leans: Soc iety for Industr ia l and App liedM athem atics, 2007: 1 027-1 035.
[ 4] L ikas A, V lassisN, V erbeek J. The g lobalK-m eans c lustering a lgor ithm [ J] . Pattern Recogn ition, 2003, 36( 2): 451-461.
[ 5] Kennedy J, Eberhart R, Sh iY. Sw arm Intelligence[M ] . [ S. .l ]: M o rgan Kaufm ann Publishers, 2001.
[ 6] de Castro L N, Von Zuben F J. Recent Deve lopm ents in B io log ically Insp ired Com puting[M ]. London: Idea Group Inc, 2004,
[ 7] Bonabeau E, DorigoM, Th?rau laz G. Sw arm Inte llig enceFrom Natura l toArtific ia l System s[M ] . [ S. .l ]: Ox fo rd Un iversity Press, 1999.
[ 8] Kennedy J, Eberhart R C. Partic le swa rm optim ization[ C] / / Proceed ings of IEEE Internationa l Conference on N eural Ne-t w orks, IV. Perth: IEEE, 1995: 1 942-1 948.
[ 9] Van derM erwe DW, Enge lbrech t A P. Data cluster ing using partic le sw arm optim ization[ C ] / / Proceed ings of IEEE Cong ress on Evolutionary Com putation 2003 ( CEC 2003), Canberra: IEEE, 2003: 215-220.
[ 10] Shubham Agraw a,l Pan ig rah i B K, M ano jKum ar T iw ar.i M ultiob jective particle sw arm a lgor ithm w ith fuzzy c lustering for elec trical pow er dispa tch[ J]. IEEE Transac tions on E vo lutiona ry Computation, 2008, 12( 5): 529-541.
[ 11] Chang Dongx ia, Zhang Xianda, Zheng Changw en. A genetic a lgo rithm w ith g ene rearrangem ent forK-means cluster ing [ J]. Pattern Recogn ition, 2009, 42( 7): 1 210-1 222.
[ 12] La i J Z C, H uang T J, Liaw Y C. A fastK-m eans cluster ing a lgo rithm using c luster center d isp lacement [ J]. Pa ttern Recogn ition, 2009, 42( 11): 2 551-2 556.

备注/Memo

备注/Memo:
基金项目: 山东省科技攻关计划项目( 2009GG10001008)、山东省软科学研究计划项目( 2009RKA285)、济南市高校院所自主创新项目 ( 200906001) . 通讯联系人: 曲建华, 讲师, 研究方向: 数据挖掘信息系统. E-mail:qujh2003@ sohu. com
更新日期/Last Update: 2013-04-08