一种基于粒子群优化的可能性C均值聚类改进方法

提出了一种基于拉子群优化的可能性c均值(Possibilistic Gmeans, PCM)聚类改进方法。该方法首先通过改进PCM算法的目标函数来计算数据模式的隶属度矩阵和聚类中心完成粒子编码，从而降低算法对初始中心的敏感，提高聚类的精度;其次，通过粒子群优化(Particle Swarm Optimization, PSO)算法对编码进行优化，以有效地克服PCM聚类算法容易导致聚类一致性和陷入局部最优解的缺点，减少算法的迭代次数。通过人造数据集和UCI数据集上的实验，表明该算法在计算复杂度、聚类精度和全局寻优能力方面表现得较为突出。

Improved Possibilistic C-means Clustering Algorithm Based on Particle Swarm Optimization

An improved possibifistic Gmeans(PCM) algorithm based on particle swarm optimization (PSO) was pre-sented. This algorithm consists of two steps; first, using the improved PCM to calculate the degree of membership ma-trix and cluster centroid to encode particles, which can low the influence of initialized centroid and improve clusteringprecision. In the second, using PSO to optimize the encoded data points, which can overcome the coincident clusters andavoid easily falling into local optimum The experimental results on the synthetic data sets and UCI data sets show thatthe proposed algorithm has less computational complexity, higher clustering precision and greater searching capability.