Utilities for manipulating correlated Gaussian random variables.
Project description
This package facilitates the creation and manipulation of arbitrarily complicated (correlated) multi-dimensional Gaussian random variables. The random variables are represented by a new data type (gvar.GVar) that can be used in arithmetic expressions and pure Python functions. Such expressions/functions create new Gaussian random variables while automatically tracking statistical correlations between the new and old variables. This data type is useful for simple error propagation, but also is heavily used by the Bayesian least-squares fitting module lsqfit.py to define priors and specify fit results, while accounting for correlations between all variables. Documentation can is in the doc/ subdirectory: see doc/html/index.html or look online at <https://gvar.readthedocs.io>.
These packages use numpy for efficient array arithmetic, and cython to compile efficient code. gvar uses automatic differentiation to track covariances through arbitrary arithmetic.
Information on how to install the components is in the INSTALLATION file.
To test the libraries try make tests. Some examples are give in the examples/ subdirectory.
gvar version numbers have the form major.minor.patch where: incompatible changes are signaled by incrementing the major version number, the minor number signals new features, and the patch number signals bug fixes.
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
| Filename, size | File type | Python version | Upload date | Hashes |
|---|---|---|---|---|
| Filename, size gvar-11.9.1.tar.gz (1.9 MB) | File type Source | Python version None | Upload date | Hashes View |
