This repository contains Python (Jupyter Notebooks), C and Shell code, which was used to generate figures in [1].
All C programs are contained within crun directory. They can be compiled
by running crun/compile-all.sh script (dependencies: GCC, GSL). After
compilation all programs can be run by executing crun/run-all.sh script
(dependencies: GNU parallel, paste).
C programs will generate multiple temporary *.series files, which will be
later combined into *.data files, which will be stored in the data
directory (it will be created as needed).
For exact details about our implementation of the noisy voter model using Gillespie method see [1]. In short, we adjust the number of agents in real time to make the simulation run faster (minutes instead of weeks).
*.data files are CSV files, which contain time series of separate runs as
rows. All runs within same file were obtained using the same parameter set.
For actual parameter set see the corresponding *.c file.
*.data files are analyzed by convert-*.ipynb notebooks
(dependencies: numpy, scipy, matplotlib). These notebooks
will convert data into *.csv files, which can be then used to produce figures
from the paper (either using matplotlib, gnuplot or any other graphing library).
Columns of the *.csv file correspond to:
- log10(time)
- log10(empirical mean)
- log10(empirical variance)
- log10(theoretical mean)
- log10(theoretical variance)
- log10(steady state mean)
- log10(steady state variance)
Feel free to copy, use and modify our code as you see fit (see the LICENSE
file for legal details). That said, referencing the paper would be appreciated.
- R. Kazakevičius, A. Kononovicius. Anomalous diffusion in nonlinear transformations of the noisy voter model. Physical Review E 103: 032154 (2021). doi: 10.1103/PhysRevE.103.032154. arXiv:2011.02927 [cond-mat.stat-mech].