Author: Dirk Hesse <herr.dirk.hesse@gmail.com>

We implement the method to estimate autocorrelation times of Monte Carlo data presented in

U. Wolff [ALPHA Collaboration], Monte Carlo errors with less errors, Comput. Phys. Commun. 156, 143 (2004) [hep-lat/0306017].

PUBLICATIONS MAKING USE OF THIS CODE MUST CITE THE PAPER.

The main objective is the following: Data coming from a Monte Carlo simulation usually suffers from autocorrelation. It is not straight-forward to estimate this autocorrelation, which is required to give robust estimates for errors. This program implements a method proposed by Wolff to estimate autocorrelations in a safe way.

Just:

$ pip install py-uwerr

should be enough to install this library from the PyPI.

This package contains code to generate correlated data, so we can conveniently demonstrate the basic functionality of the code in a short example:

>>> from puwr import tauint, correlated_data
>>> correlated_data(2, 10)
[[array([ 1.02833043,  1.08615234,  1.16421776,  1.15975754,
          1.23046603,  1.13941114,  1.1485227 ,  1.13464388,
          1.12461557,  1.15413354])]]
>>> mean, delta, tint, d_tint = tauint(correlated_data(10, 200), 0)
>>> print "mean = {0} +/- {1}".format(mean, delta)
mean = 1.42726267057 +/- 0.03013853
>>> print "tau_int = {0} +/- {1}".format(tint, d_tint)
tau_int = 9.89344869217 +/- 4.10466090332

The data is expected to be in the format data[observable][replicum][measurement]. See the documentation that comes with this code for more information.

See LICENSE file.