基于PSO的模糊聚类算法

Fuzzy clustering algorithm based on PSO

提出了一种基于模糊C-均值算法和粒子群算法的混合聚类算法。该算法结合PSO的全局搜索和FCM局部搜索的特点,将PSO优化聚类结果作为后续FCM算法的初始值,有效地克服了FCM对初始值敏感、易陷入局部最优和PSO算法局部搜索较弱的问题,同时增强了跳出局部最优的能力。实验表明,新算法得到的目标函数值更小,并能减小分类错误率,聚类效果优于单一使用FCM或PSO。

A new hybrid clustering algorithm based on particle swarm optimization and FCM algorithm is proposed. By incorporating the local and global search and taking the clustering result of PSO as the initialized value of the FCM, the algorithm eliminates FCM trapped local optimum and being sensitive to initial value effectively, and solves weaker local search of PSO. The ability of breaking away from the local optimum is improved by the new algorithm. The experimental results show that new algorithm not only has better goal function value but also reduces the classification error rate. The clustering performances are better than those of only using the FCM or the PSO.