myFitter - Maximum Likelihood Fits in C++ and Python
myFitter is a library for maximum likelihood fits and the numerical computation of p-values in likelihood ratio tests. It has been implemented independently in C++ and Python.
In some situations the statistical significance (p-value) of a likelihood ratio test can not be computed analytically. This is, for example, the case when some parameters of the model are only allowed to float within a certain range or when the models being compared are not nested, meaning that neither model can be obtained from the other by fixing some of its parameters.
myFitter implements a method for efficient numerical computation of p-values. This method is also applicable in the case of non-nested models. It also provides a convenient framework for performing combined fits of data from different experiments.
The algorithms implemented in myFitter are discussed in
If you use myFitter for a scientific publication, please cite this article. In addition, please follow the citation guidelines of the Dvegas package, which myFitter links to.
The python version is still at an early stage of development and currently lacks complete documentation. A basic and fairly pedagogical introduction was given in a software tutorial at the YETI school 2014 in Durham. The material for the tutorial can be found here.
Please send bugreports to Martin Wiebusch