There are already two other Gaussian Process packages in Julia, however their feature are quite orthogonal. They are roughly compared here: AGP.jl stands for AugmentedGaussianProcesses.jl and GP.jl for GaussianProcesses.jl
| Likelihood | AGP.jl | Stheno.jl | GP.jl |
|---|
| Gaussian | ✓ | ✓ (multi-input/multi-output) | ✓ |
| Student-T | ✓ | ✖ | ✓ |
| Bernoulli | ✓ (Logistic) | ✖ | ✓ (Probit) |
| Bayesian-SVM | ✓ | ✖ | ✖ |
| Poisson | ✓ | ✖ | ✓ |
| NegativeBinomial | ✓ | ✖ | ✖ |
| Exponential | ✖ | ✖ | ✓ |
| MultiClass | ✓ | ✖ | ✖ |
| Inference | AGP.jl | Stheno.jl | GP.jl |
|---|
| Analytic (Gaussian only) | ✓ | ✓ | ✓ |
| Variational Inference | ✓ (Analytic and Num. Appr.) | ✖ | ✖ |
| Gibbs-Sampling | ✓ | ✖ | ✖ |
| MCMC | ✖ | ✖ | ✓ |
| Expec. Propag. | ✖ | ✖ | ✖ |
| Kernel | AGP.jl | Stheno.jl | GP.jl |
|---|
| RBF/Squared Exponential | ✓ | ✓ | ✓ |
| Matern | ✓ | ✖ | ✓ |
| Const | ✖ | ✓ | ✓ |
| Linear | ✖ | ✓ | ✓ |
| Poly | ✖ | ✓ | ✓ |
| Periodic | ✖ | ✖ | ✓ |
| Exponentiated Quadratic | ✖ | ✓ | ✖ |
| Rational Quadratic | ✖ | ✓ | ✓ |
| Wiener | ✖ | ✓ | ✖ |
| Sum of kernel | ✖ | ✖ | ✓ |
| Product of kernels | ✖ | ✖ | ✓ |
Note that the kernels will be defered to MLKernels.jl in the future.
| Feature | AGP.jl | Stheno.jl | GP.jl |
|---|
| Sparse GP | ✓ | ✖ | ✓ |
| Custom prior Mean | ✓ | ✓ | ✓ |
| Hyperparam. Opt. | ✓ | ? | ✓ |
| MultiOutput | ✖ | ✓ | ✖ |