Date Added: Jan 2011
This paper studies a clustering algorithm based on information visulization. In this algorithm, through a Non-Linear Programming mapping (NLP), some high-dimensional and complicated feature data is transformed into low-dimensional feature data, such as one, two and three dimensionality. Its main aim is that the geometry image in high-dimensional space is mapped into one, two and three dimensional image in low-dimensional space, and the inherent data "Structure" is approximately preserved after mapping. The simulated results show that the algorithm presented here is feasible and effective with direct observation and image et al.. It describes well nonlinear character for high-dimensional feature data.