摘要
文章提出一种基于粒子群方法的投影寻踪聚类算法.该算法利用粒子群算法的全局寻优能力搜索投影方向,以及利用投影寻踪算法将高维数据投影到低维,从而能有效地揭示多因素复杂数据的线性和非线性结构的特点,得到较直观的聚类模式.通过两个实例,验证了算法的可行性和有效性.
A PSO-based projection pursuit clustering algorithm is put forward in this paper. The algo- rithm can find the projection direction by making use of global searching ability of PSO, and the projection pursuit algorithm to project the data from high dimension to low dimension, the linear and nonlinear char-acter of multifactor and complex datum can efficiently be showed. Two experiments demonstrate the algo-rithm is practicable and effective.
出处
《曲阜师范大学学报（自然科学版）》
CAS
2008年第2期43-46,共4页
Journal of Qufu Normal University(Natural Science)
基金
国家自然科学基金资助项目(70471049)
关键词
投影寻踪
粒子群方法
聚类
projection pursuit
particle swarm optimization
clustering Algorithm
