基于粒子群优化的直觉模糊核聚类算法研究

doi:10.11959/j.issn.1000-436x.2015099

摘要/Abstract

摘要：

针对现有基于核方法的直觉模糊聚类算法对初始值敏感、收敛速度慢等缺陷，利用粒子群优化算法全局搜索能力强、收敛速度快的优势，对直觉模糊核聚类算法的初始聚类中心进行优化，并提出了一种基于粒子群优化的直觉模糊核聚类算法。该算法在提升聚类性能的同时，有效增强了算法的收敛速度。在实验阶段，采用4组标准数据集对该算法进行了分类实验及有效性测试，并将其与模糊c均值聚类算法及直觉模糊c均值聚类算法的分类效果及运行时间进行对比，实验结果充分表明了该算法的有效性及优越性。

关键词: 直觉模糊集, 核方法, 模糊聚类, 粒子群优化

Abstract:

The intuitionistic fuzzy kernel c-means clustering algorithm has several problems such as sensitivity to the ini-tial value,low convergence speed,etc.To overcome these shortages,the particle swarm optimization (PSO) algorithm with powerful ability of global search and quick convergence rate is applied to intuitionistic fuzzy clustering.Firstly,PSO is used to optimize the initial clustering centers.Then,the approach of intuitionistic fuzzy kernel clustering based on PSO,namely PS-IFKCM,is proposed.This algorithm can enhance both the clustering ability and the convergence speed.Fi-nally,experiments based on four measured datasets are carried out to illustrate the performance of the proposed method.Compared with results from FCM and IFKCM,PS-IFKCM is of great efficiency for classification.

Key words: intuitionistic fuzzy sets, kernel method, fuzzy clustering, particle swarm optimization

余晓东,雷英杰,岳韶华,王睿. 基于粒子群优化的直觉模糊核聚类算法研究[J]. 通信学报, 2015, 36(5): 74-80.

UXiao-dong Y,EIYing-jie L,UEShao-hua Y,ANGRui W. Research on PSO-based intuitionistic fuzzy kernel clustering algorithm[J]. Journal on Communications, 2015, 36(5): 74-80.

图/表 6

图1

图2

表1

图3

图4

表2

参考文献 25

[1]	RUSPINI E H . A new approach to clustering[J]. Information and Control, 1969,15(1): 22-32.
[2]	CECCARELLI M H , MARATEA A . Improving fuzzy clustering of biological data by metric learning with side information[J]. Int'l Journal of Approximate Reasoning, 2008,47(1): 45-57.
[3]	张洪美，徐泽水，陈琦．直觉模糊集的聚类方法研究[J]．控制与决策， 2007,22(8):882-888. ZHANG H M , XU Z S , CHEN Q . On clustering approach to intuitionistic fuzzy sets[J]. Control and Decision, 2007,22(8):882-888.
[4]	DUMN J C . A graph theoretic analysis of pattern classification via tamura's fuzzy relation[J]. IEEE Trans on Fuzzy System, 1974,4(3): 310-313.
[5]	BEZDEK J C . Pattern Recognition with Fuzzy Objective Function Algorithms[M]. New York Plenum Press, 1981.
[6]	XUE Z K , LI S Y . A multi-model modeling approach to MIMO nonlinear systems[J]. Acta Electronic Sinica, 2005,33(1): 52-56.
[7]	雷英杰，王宝树，苗启广．直觉模糊关系及其合成运算[J]．系统工程理论与实践， 2005,25(2):113-118,133. LEI Y J , WANG B S , MIAO Q G . On the intuitionistic fuzzy relations with compositional operations[J]. Systems Engineering Theory and practice, 2005,25(2):113-118,133.
[8]	郑寇全，雷英杰，王睿等．基于 IFSTR 的抽象化空间推理方法[J]．系统工程与电子技术， 2013,35(3):651-654. ZHENG K Q , LEI Y J , WANG R , et al. The method for abstract spatial reasoning based on IFSTR[J]. Systems Engineering and Electronics, 2013,35(3):651-654.
[9]	樊雷，雷英杰．直觉模糊统计决策与判决[J]．控制与决策， 2011,26(3):357-362. FAN L , LEI Y J . Intuitionistic fuzzy statistic adjudging and decision-making[J]. Control and Decision, 2011,26(3):357-362.
[10]	贺正洪，雷英杰．直觉模糊 c 均值聚类算法研究[J]．控制与决策， 2011,26(6):847-850,856. HE Z H , LEI Y J . Research on intuitionistic fuzzy c-means clustering algorithm[J]. Control and Decision, 2011,26(6):847-850,856.
[11]	申晓勇，雷英杰，李进等．基于目标函数的直觉模糊集合数据的聚类方法[J]．系统工程与电子技术， 2009,31(11):2732-2735. SHEN X Y , LEI Y J , LI J , et al. Clustering technique to intuitionistic fuzzy sets data based on objective function[J]. Systems Engineering and Electronics, 2009,31(11):2732-2735.
[12]	CORTES C , VAPNIK V . Support-vector networks[J]. Machine Learning, 1995,20(3): 273-297.
[13]	CHEN S C , ZHANG D Q . A novel kernelized fuzzy c-means algo-rithm with application in medical image segmentation[J]. Artificial Intelligence in Medicine, 2004,32(1): 37-50.
[14]	CAMASTRA F , VERRI A . A novel kernel method for clustering[J]. IEEE Trans on Pattern Analysis and Machine Intelligence, 2005,27(5): 801-805.
[15]	FILIPPONE M , CAMASTRA F , MASULLI F , et al. A survey of kernel and spectral methods for clustering[J]. Pattern Recognition, 2008,41(1): 176-190.
[16]	GIROLAMI M . Mercer kernel based clustering in feature space[J]. IEEE Trans on Neural Networks, 2002,13(3): 780-784.
[17]	张莉，周伟达，焦李成．核聚类算法[J]．计算机学报， 2002,25(6):587-590. ZHANG L , ZHOU W D , JIAO L C . Kernel clustering algorithm[J]. CHINESE J.COMPUTERS, 2002,25(6):587-590.
[18]	范成礼，邢清华，付强等．基于直觉模糊核聚类的弹道中段目标识别方法[J]．系统工程与电子技术， 2013,35(7):1362-1367. FAN C L , XING Q H , FU Q , et al. Technique for target recognition in ballistic midcourse based on intuitionistic fuzzy kernel clustering[J]. Systems Engineering and Electronics, 2013,35(7):1362-1367.
[19]	曲福恒，马驷良，胡雅婷．一种基于核的模糊聚类算法[J]．吉林大学学报， 2008,46(6):1137-1141. QU F H , MA S L , HU Y T . A kernel based fuzzy clustering algorithm[J]. Journal of Jilin University, 2008,46(6):1137-1141.
[20]	雷阳，雷英杰，周创明．基于直觉模糊核匹配追踪的目标识别方法[J]．电子学报， 2011,39(6):1441-1446. LEI Y , LEI Y J , ZHOU C M . Techniques for target recognition based on intuitionistic fuzzy kernel matching pursuit[J]. Acta Electronica Sinica, 2011,39(6):1441-1446.
[21]	雷阳，孔韦韦，雷英杰．基于直觉模糊 c 均值聚类核匹配追踪的弹道中段目标识别方法[J]．通信学报， 2012,33(11):136-143. LEI Y , KONG W W , LEI Y J . Technique for target recognition based on intuitionistic fuzzy c-means clustering and kernel matching pursuit[J]. Acta Electronica Sinica, 2012,33(11):136-143.
[22]	KENNEDY J , EBERHART R C . Particle swarm optimization[A]. Proc of IEEE Internal Conference on Neural Networks Perth[C]. 1995. 1942-1948.
[23]	RANA S , JASOLA S , KUMAR R . A review on particle swarm opti-mization algorithms and their applications to data clustering[J]. Artificial Intelligence Review, 2011,35(3): 211-222.
[3]	王纵虎，刘志镜，陈东辉．一种改进的粒子群优化快速聚类算法[J]．西安电子科技大学学报， 2012,39(5):61-65,78. WANG Z H , LIU Z J , CHEN D H . Improved PSO-based fast clustering algorithm[J]. Journal of Xidian University, 2012,39(5):61-65,78.
[25]	于德亮，唐海燕，丁宝等．基于粒子群优化模糊核聚类的电梯群交通模式识别[J]．哈尔滨工业大学学报， 2012,44(10):84-88. YU D L , TANG H Y , DING B , et al. Elevator traffic mode identification with kernel fuzzy clustering based on particle swarm optimization[J]. Journal of Harbin Institute of Technology, 2012,44(10):84-88.

数据集	FCM	IFKCM	PS-IFKCM
Iris	92.76	94.74	96.91
Wine	92.36	95.37	96.73
Wisc	93.58	95.46	96.89
Motorcycle	93.26	94.88	97.07

数据集	FCM	IFKCM	PS-IFKCM
Iris	0.106	6.256	2.025
Wine	0.289	9.134	3.131
Wisc	0.365	9.789	3.541
Motorcycle	0101	4.378	1.416