摘要
建立了聚类分析问题的数学优化模型,提出了一种新的粒子群算法解决聚类问题。对基本粒子群优化算法作了改进,思路是将K-均值方法的结果作为一个粒子和利用新的分类中心调整粒子位置。对Iris植物样本数据的测试结果表明:4种粒子群算法的效果都比较好,特别是第3种改进的粒子群算法的效果更好,粒子群优化聚类技术很有潜力．
An optimization mathematical model of the clustering problem is given. A new approach using particle swarm optimization （PSO） is put forward to solve the clustering problem. The basic PSO algorithm is extended by using K-means clustering to seed the initial swarm and the particle position is adjusted according to new cluster center vectors. The algorithms are evaluated on Iris plants database. Results show that all the 4 particle swarm optimization algorithms are proved to be effective and especially the third modified PSO algorithm is proved to be more effective. It also shows that PSO clustering techniques have much potential.
出处
《南京航空航天大学学报》
EI
CAS
CSCD
北大核心
2006年第B07期62-65,共4页
Journal of Nanjing University of Aeronautics & Astronautics
关键词
粒子群算法
聚类问题
优化
particle swarm algorithm
clustering problem
optimization
