Gaussian Processes for Machine Learning (GPML) Toolbox

• DOI: 10.5555/1756006.1953029
• Publication Date: 2010-03-01
• Authors: Carl Edward Rasmu...; Hannes Nickisch
• Keywords: Expectation propagation; Laplace's method; Toolbox

The GPML toolbox provides a wide range of functionality for Gaussian process (GP) inference and prediction. GPs are specified by mean covariance functions; we offer library simple functions mechanisms to compose more complex ones. Several likelihood supported including heavy-tailed regression as well others suitable classification. Finally, methods is provided, exact variational inference, Expectation Propagation, Laplace's method dealing with non-Gaussian likelihoods FITC large tasks.