https://github.com/GPflow/GPflow
Tip revision: 71eadd1d17286e63ed313e4607553beb6b3b666b authored by vdutor on 20 December 2018, 11:40:38 UTC
Start Shared Mixed Mok
Start Shared Mixed Mok
Tip revision: 71eadd1
model_options.rst
========================
The six core models of GPflow
========================
The following table summarizes the six core model options in GPflow.
+----------------------+--------------------------+----------------------------+-----------------------------+
| | Gaussian | Non-Gaussian (variational) | Non-Gaussian |
| | Likelihood | | (MCMC) |
+======================+==========================+============================+=============================+
| Full covariance | :class:`gpflow.models.GPR` | :class:`gpflow.models.VGP` | :class:`gpflow.models.GPMC` |
+----------------------+--------------------------+----------------------------+-----------------------------+
| Sparse approximation | :class:`gpflow.models.SGPR`| :class:`gpflow.models.SVGP` | :class:`gpflow.models.SGPMC` |
+----------------------+--------------------------+----------------------------+-----------------------------+
The GPLVM which adds latent variables is also included (`notebook <notebooks/GPLVM.html>`_).
GP Regression
-------------
.. automodule:: gpflow.models
.. autoclass:: gpflow.models.GPR
Sparse GP Regression
--------------------
See also the documentation of the `derivation <notebooks/SGPR_notes.html>`_.
.. automodule:: gpflow.models
.. autoclass:: gpflow.models.SGPR
Variational Gaussian Approximation
----------------------------------
See also the documentation of the `derivation <notebooks/VGP_notes.html>`_.
.. automodule:: gpflow.models
.. autoclass:: gpflow.models.VGP
Sparse Variational Gaussian Approximation
-----------------------------------------
.. automodule:: gpflow.models
.. autoclass:: gpflow.models.SVGP
Markov Chain Monte Carlo
------------------------
.. automodule:: gpflow.models
.. autoclass:: gpflow.models.GPMC
Sparse Markov Chain Monte Carlo
-------------------------------
.. automodule:: gpflow.models
.. autoclass:: gpflow.models.SGPMC