A modified approximate method based on Gaussian radial basis functions

摘要

A modified approximate method based on Gaussian radial basis functions is presented, and this method is suitable for scattered data approximation, moreover, this method is not too sensitive to outliers. An iterative procedure for computing the unknown coefficient is demonstrated and analyzed. Numerical examples for curve and surface fitting in different cases are given to illustrate that this method is accurate, effective and feasible, especially in the presence of noise and outliers.