Gaussian Mixture Regression
This is a MatLab implementation of a regression algorithm based on a Gaussian Mixture Model. GMR uses an EM algorithm to estimate a joing GMM among the input and output variables. For prediction it uses a weighted conditional Gaussian distribution based on the trained GMM model. It can be used for one or multiple outputs.
Run ScriptDemo.m to see an example.
This code depends on the NetLab toolbox. It is free available from: http://www.aston.ac.uk/eas/research/groups/ncrg/resources/netlab/downloads/