Abstract
In this paper, we will propose sparse algorithm for RBF approximation. RBF is good approximation for both local accuracy and global follow up. However, when the number of basis rises, it sometimes need too much time consuming to calculate its response. Especially, when we would like to optimize radius, it took too much time. When we use RBF for control problem, these quickness of response is strongly required. We will show the basic rule of sparse and show its effectivenss.
