BayesicFitting is a toolbox for the fitting of models to data.

Data in this context means a set of (measured) points x and y. The model provides some (mathematical) relation between the x and y. Fitting adapts the model such that certain criteria are optimized.

The BayesicFitting toolbox also answers the question whether one model fits the data better than another. Especially this latter aspect raises Bayesian fitting above other model fitting procedures.

BayesicFitting consists of more than 100 Python classes, of which a third are model classes. Another third are fitters in one guise or another and miscelleaneous stuff. The remaining third is needed for Nested Sampling, a novel way to solve Bayesian inference problems, developed by John Skilling and David MacKay.

The package is stored at github and pypi. The easiest way to install it is via

> pip install BayesicFitting
See also the Quick section.

This site is dedicated to documentation. It contains manuals, design documents and detailed documentation on all classes and methods. The glossary defines all terms used throughout the package. In case of troubles during the fitting process the Help document provides help, also on restrictions on data and models. The fitting process is one part mathematics and one part art: the art of dealing with finite precision and time availble in our computers.

The pictures on the right side are a few examples of what can achieved with the toolbox. More extensive information on them can be found in the Examples section. More examples are in the directory examples at github.