Gaussian Processes

Supports log-likelihood optimization, Hamiltonian Monte Carlo, elliptical slice sampling, and variational inference, catering to diverse Bayesian inference needs as detailed in the functionality section.

Handles non-Gaussian observations through Monte Carlo sampling from the intractable posterior, enabling modeling beyond standard Gaussian likelihoods.

Includes various mean, kernel, and likelihood functions, allowing users to customize models for specific data characteristics, as highlighted in the documentation.

Integrates with ScikitLearn.jl for hyperparameter tuning and cross-validation, enhancing machine learning workflows as shown in the provided example notebook.