Gaussian process regression using user inputs.
demogp_simp
demogp_simp - is an interactive demo of the Gaussian process
regression.
The code is simple, designed to provide a visual demonstration tool where the Gaussian Process learns the model as new data is presented to it.
Two windows show up: the left one is for adding training data and in the right window samples from the posterior are drawn.
The left figure shows the mean of the posterior process, two standard deviation of the posterior and two standard deviation of the predictive distribution (the latter one is displayed assuming Gaussian likelihood model).
Data addition is ended if return is pressed instead a mouse click when adding an input.
The parameters of the Gaussian process - the type of kernel and the likelihood function - have to be set in the source code.
Copyright (c) Lehel Csató (2001-2004)