摘要
针对常规动态聚类方法对初始聚类中心的敏感性和容易陷入局部最优的缺点等问题,提出了基于二阶段微粒群优化模糊C-均值算法(TPSOFCM),并将此算法与现有的模糊C-均值聚类算法和基于多阶段的模糊C-均值算法进行比较。该算法对Iris数据进行聚类,计算结果表明:该算法有较好的全局收敛性,不仅能有效地克服传统的模糊C-均值算法易陷入局部最优解的缺点,而且全局收敛能力优于模糊C-均值聚类算法和基于多阶段的模糊C-均值算法。
After analyzing the disadvantages of Fuzzy C - means Clustering algorithm sensitive to the initial value and easy to fall into the local optimization, this paper proposes a new clustering method based on Two - stage Particle Swarm Optimization Fuzzy C - Means Algorithm, and compares it with FCM algorithm and MFCM algorithm. Numerical experiment is made on the Iris data. The research indicates that the algorithm not only avoids the local optimization, but also has greater searching capability than the existing FCM algorithm and MFCM algorithm.
出处
《陕西理工学院学报(自然科学版)》
2007年第1期77-80,共4页
Journal of Shananxi University of Technology:Natural Science Edition
关键词
全局最优
聚类分析
微粒群优化算法
隶属函数
global optimization
clustering analysis
Particle Swarm Optimization algorithm
membership function.