THunc
a tool for computing theory uncertainties from missing higher orders
Author: Marco Bonvini
Developer: Marco Bonvini
THunc is a code that computes a probability distribution for a physical observable based on its perturbative expansion.
From the probability distribution a number of statistical estimators (mean, mode, median, standard deviation, degree of belief intervals) are also provided.
The calculation of the probability distribution is based on some assumptions, that are encoded into a model.
The models available in the code are
- Geometric behaviour model (assumes the expansion is bounded by a geometric series)
- Scale variation model (uses information on the scale dependence to put constraints on the next orders) [QCD only]
- Custom model (a customisable model based on user-defined assumptions)
In all cases one can also look for a sign pattern through the flag -P.
Scale independent results can be obtained using the flag -s [QCD only].
Use the -h option for help.
To set the process and/or to define a custom model, open the example_QCD.cc file and check its first part.
Obviously, the code must be recompiled after changes.
Download
Latest version: THunc v1.1
Instructions: Download the tarballs above, or wget them directly:
> wget https://www.roma1.infn.it/~bonvini/THunc/downloads/THunc.v1.1.tgz
then untar, make and run
> tar xzf THunc.v1.1.tgz
> cd THunc
> make -j
> ./examples/example_QCD -r 0.5
> gnuplot plots/gp_scale_dependence.gp
> gnuplot plots/gp_distributions.gp
> gnuplot plots/gp_summary.gp
Version history
- 11-09-2020: THunc v1.1
Fixed a memory leak (custom model now faster). Fixed compiler problems on Linux.
- 29-06-2020: THunc v1.0
Initial release
Dependencies
THunc depends on the following libraries:
Citation Policy
When you use the THunc code, please cite this webpage and the original publications:
\cite{Bonvini:2020xeo}
Contacts
For any problem, request or suggestion, contact me
- marco.bonvini [at] roma1.infn.it