【摘要】 聚类算法在数据分析及数据挖掘等许多领域有广泛应用,在聚类方法中引入一种新的距离度量标准替代传统的Eu-clidean距离度量标准以提高其健壮性,并在此基础上提出基于粒子群算法(Particle Swarm Optimization,简称PSO)的聚类方法和基于量子行为的微粒群优化算法(Quantum-behaved Particle Swarm Optimization,简称QPSO)的聚类方法,然后将两种聚类方法应用于图像分割。实验结果表明,基于QPSO的聚类方法性能优于基于PSO的聚类方法。更多还原

【Abstract】 Clustering algorithm has got extensive applications in many fields, for example data analysis and data dig. A new distance metric is proposed to replace Euclidean distance for improving robust. Based on it Particle Swarm Optimization (PSO) clustering algo-rithm and Quantum-behaved Particle Swarm Optimization (QPSO) clustering algorithm are introduced. Then two clustering algorithm are used in image color segmentation. The experimental results show that QPSO clustering algorithm has better capability than PSO clustering algorithm.更多还原