Radial basis functions (HTML5)

Gaussian kernel Multiquadric Inverse multiquadric Inverse quadratic Polyharmonic spline Inverse distance weighting	Add Point	Remove Point	Reset Points	Interpolate Points	Draw sum
$${}$$					1.0
					k = 1 k = 2 k = 3 k = 4 k = 5

A set of scattered data points \(p_i\) with scalar values \(f_i\) is approximated as a weighted sum of \(N\) radial functions \(\varphi\). Each radial function \(\varphi(r)\) is centered around a data point \(p_i\) and decreases quickly with increasing distance to \(p_i\). Click on "interpolate points" to find the weights \(w_i\) such that \(f(p_j)=f_j\) for \(j=1,\ldots,N\). A system of \(N\) linear equations is then solved using Gaussian elimination.

Applet created by Johannes Kehrer for educational purpose.