Manifold Clustering Based on Differential Evolution
Clustering is an important technique in data mining and business intelligence.Although there are many proposed algorithms and applications for clustering,one of the basic assumptions is that the data is distributed in a smooth space,the Euclidean space Rn for example.On the other hand,more and more complex data types emerge frequently in the form of manifold,approximately a curved space,locally alike smooth space.The purpose of this paper is to propose new clustering technique on manifolds,that is,on curved spaces.This is achieved mainly with the help of tangent spaces that are determined by manifold learning.We embed a new searching algorithm based differential evolution (DE) which proves to be a simple optimization algorithm effective for real-valued problems.We present a simple convergence analysis with a design of experimental framework.
Cluster analysis manifold data mining density-based clustering gravity-based clustering Differential Evolution tangent space
Xiyu Liu Liandi Jiang Jianping Zhang
School of Management and Economics,Shandong Normal University,Jinan,P.R.China
国际会议
长沙
英文
173-184
2008-10-28（万方平台首次上网日期，不代表论文的发表时间）