Skip to main content

mkgp

These classes and routines were developed by Aaron Ho, and this project repository was started in 2017. The underlying mathematics was founded on the book, “Gaussian Process for Machine Learning”, C.E. Rasmussen, C.K.I. Williams (2006).

When using this package in any research work, please cite: A. Ho et al 2019 Nucl. Fusion 59 056007, DOI: 10.1088/1741-4326/ab065a

Note that the package has been renamed from GPR1D -> mkgp in v3.0.0.

Installing the mkgp program

Installation is mandatory for this package!

For first time users, it is strongly recommended to use the GUI developed for this Python package. To obtain the Python package dependencies needed to use this capability, install this package by using the following on the command line:

pip install [--user] mkgp[gui]

Use the --user flag if you do not have root access on the system that you are working on. If you have already cloned the repository, enter the top level of the repository directory and use the following instead:

pip install [--user] -e .[gui]

Removal of the [gui] portion will no longer check for the pyqt5 and matplotlib packages needed for this functionality. However, these packages are not crucial for the base classes and algorithms.

To test the installation, execute the command line script:

mkgp_1d_demo

This demonstration benefits from having matplotlib installed, but is not required.

Documentation

Documentation of the equations used in the algorithm, along with the available kernels and optimizers, can be found in docs/. Documentation of the mkgp module can be found on GitLab pages

Using the gpr1d program

For those who wish to include the functionality of this package into their own Python scripts, a sample script is provided in src/mkgp/scripts/demo.py. The basic syntax used to create kernels, select optimizers, and perform the GP regression fits are outlined there.

For any questions or to report bugs, please do so through the proper channels in the GitLab repository.

Important note for users!

The following runtime warnings are common within this routine:

RuntimeWarning: overflow encountered in double_scalars
RuntimeWarning: invalid value encountered in true_divide
RuntimeWarning: invalid value encountered in sqrt

They are filtered out by default but may reappear if verbosity settings are modified. They normally occur when using the kernel restarts option (as in the demo) and do not necessarily mean that the final returned fit is poor.

Plotting the returned fit and errors is the recommended way to check its quality. The log-marginal-likelihood metric can also be used, but is only valuable when comparing different fits of the same data, i.e. its absolute value is meaningless.

From v1.1.1, the adjusted R2 and pseudo R2 metrics are now available. The adjusted R2 metric provides a measure of how close the fit is to the input data points. The pseudo R2 provides a measure of this closeness accounting for the input data uncertainties.

Metadata

Release files for mkgp 3.1.5

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for mkgp 3.1.5
File Size Uploaded Provenance Actions
mkgp-3.1.5.tar.gz 808.0 kB Without provenance Details for mkgp-3.1.5.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for mkgp 3.1.5
File Interpreter ABI Platform Actions
mkgp-3.1.5-py3-none-any.whl Python 3 none any Details for mkgp-3.1.5-py3-none-any.whl

Total release size: 878.2 kB

Release files / mkgp-3.1.5.tar.gz

Download URL mkgp-3.1.5.tar.gz
Size 808.0 kB
Tags Source
SHA-256 checksum
How to use checksums
1189cda8a97b190b659fe73183c0433178ddfeb3372586ba7ae35b65b830a0c1
BLAKE2b-256 checksum
How to use checksums
df6f169c03a189446ecaa956f8969e215cb4feadc95d3be18040cf7f7f5d70db
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.12.13

Release files / mkgp-3.1.5-py3-none-any.whl

Download URL mkgp-3.1.5-py3-none-any.whl
Size 70.2 kB
Tags Python 3
SHA-256 checksum
How to use checksums
e2a3e9ade83495925565ded52ac18048eee3d5251a09b426ebd1cfcbf34a987f
BLAKE2b-256 checksum
How to use checksums
c945d04333d1990a5cb4761f2d96b36ebdac95ccab42c104f5e2ac9b6ff347bd
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.12.13

Security for mkgp 3.1.5

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page