Gaussian process and RBF interpolation implemented in Python. Based on C. M. Bishop's book Pattern Recognition and Machine Learning (2006).
Given a random set of samples x, and their respective values y, which arise from some function f(x), we can construct a kernel and predict for values of x not present in the input. When trying to predict values for x with gaussian process the kernel needs to be constructed with great care, often a mix of different kernel functions gives the best performance.
A set of 2D points uniformly generated using f(x,y), between the range of {-1.0, 1.0}, are interpolated using RBF interpolation.