Computation of Kemeny's constant using censored Markov Chains.
This is the code accompanying the paper "On Kemeny’s constant and stochastic complement" by
The preprint is available on arXiv and - if you use this code - can be cited as:
@misc{bini2023kemenys,
title={On Kemeny's constant and stochastic complement},
author={Dario Andrea Bini and Fabio Durastante and Sooyeong Kim and Beatrice Meini},
year={2023},
eprint={2312.13201},
archivePrefix={arXiv},
primaryClass={math.NA}
}
To get the code with all the submodules you can use the command:
git clone --recurse-submodules git@github.com:Cirdans-Home/Kemeny-and-Conquer.git
Some of the tests make use of the stochastic trace estimator called Hutch++. The repository with related code is added as a git submodule. If you use that version of the algorithm for estimating the Kemeny constant, you should also cite the original article:
@incollection {MR4537958,
AUTHOR = {Meyer, Raphael A. and Musco, Cameron and Musco, Christopher
and Woodruff, David P.},
TITLE = {Hutch++: optimal stochastic trace estimation},
BOOKTITLE = {Symposium on {S}implicity in {A}lgorithms ({SOSA})},
PAGES = {142--155},
PUBLISHER = {[Society for Industrial and Applied Mathematics (SIAM)],
Philadelphia, PA},
YEAR = {2021},
ISBN = {978-1-61197-649-6},
MRCLASS = {68W20 (15A15)},
MRNUMBER = {4537958},
}
One of the implemented version of the recursive algorithm for the approximation of the Kemeny constant uses the hm-toolbox by Massei, Robol and Kressner. If you use that version of the algorithm for estmating the Kemeny constant, you should also cite the original article:
@article {MR4082285,
AUTHOR = {Massei, Stefano and Robol, Leonardo and Kressner, Daniel},
TITLE = {hm-toolbox: {MATLAB} software for {HODLR} and {HSS} matrices},
JOURNAL = {SIAM J. Sci. Comput.},
FJOURNAL = {SIAM Journal on Scientific Computing},
VOLUME = {42},
YEAR = {2020},
NUMBER = {2},
PAGES = {C43--C68},
ISSN = {1064-8275,1095-7197},
MRCLASS = {65F99 (15A03 65F05 65F08 65F10 65F50 65Y15)},
MRNUMBER = {4082285},
MRREVIEWER = {Bruno\ Carpentieri},
DOI = {10.1137/19M1288048},
URL = {https://doi.org/10.1137/19M1288048},
}