Computer Engineering and Applications ›› 2012, Vol. 48 ›› Issue (35): 151-155.

Previous Articles     Next Articles

Research of improved fuzzy C-means algorithm based on quantum-behavior particle swarm optimization

LI Yin, MAO Li, XU Wenbo   

  1. Key Lab of Advanced Process Control for Light Industry(Ministry of Education), School of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China
  • Online:2012-12-11 Published:2012-12-21

量子粒子群优化改进的模糊C均值聚类算法

李  引,毛  力,须文波   

  1. 江南大学 物联网工程学院,轻工过程先进控制教育部重点实验室,江苏 无锡 214122

Abstract: Fuzzy C-Means(FCM) clustering algorithm has the shortcomings of being sensitive to the initial cluster centers and being trapped by local optima, To resolve two disadvantages, this paper proposes a novel clustering method using Quantum-behavior Particle Swarm Optimization(AQPSO) to optimize the improved FCM clustering algorithm(AF-AQ-AF), AQPSO algorithm is introduced based on a new metric standard which can lower the influence of initialized data points, quickly converge to the optimal solution and improve the global search ability. Data experimental results show the proposed algorithm avoids entering local minimum, enhances the convergence rate and gets a better result of clustering.

Key words: cluster analysis, Fuzzy C-Means(FCM), Quantum-behavior Particle Swarm Optimization(QPSO), new metric standard

摘要: 针对模糊C-均值(FCM)聚类算法对初始聚类中心选择敏感,易陷入局部最优的问题,提出一种量子粒子群优化改进的模糊C均值聚类算法。该算法引入的基于新距离标准的量子粒子群(AQPSO)算法不仅可以降低初始点敏感度,较快地收敛到最优解,而且能够提高全局搜索能力。仿真实验证明,该融合算法在摆脱局部最优区域,保证收敛速度同时使得聚类效果较好。

关键词: 聚类分析, 模糊C-均值(FCM), 量子粒子群(QPSO), 新距离标准