Skip to content

NAU-CCL/PMCMC

Repository files navigation

Estimating Dynamic Transmission Rates with Black-Karasinski Process in Stochastic SIHR models using particle MCMC

This codebase can be used to replicate our results from the paper.

Setup

It is recomennded to use a python virtual environment to setup the python environment. This can be done easily in vscode or from the command line.

Command line

pip install virtualenv (if you don't already have virtualenv installed)
virtualenv venv to create your new environment (called 'venv' here)
source venv/bin/activate to enter the virtual environment
pip install -r requirements.txt to install the requirements in the current environment

VSCode

Press ctrl-shift-p, cmd-shift-p on mac. Choose Python:Create Environment, then follow the steps.

Folder Structure

The examples subdirectory features two files designed to serve as a simple example of the code. The first is to generate a synthetic dataset, the second performs PMCMC on the data.

Note

The .sh files can be run using sbatch job_file.sh (job_file is a placeholder for the filename) The bash scripts are designed to be run in a SLURM HPC environment. The python scripts can be run on their own, but PMCMC is quite computationally intensive.

Experiment_1 contains the files required to replicate the results and figures from Experiment 1 in the paper. To replicate the results run exp_1_job.sh and once finished, viz.ipynb to create the plots.

Experiment 2 contains the job file and the python file used to generate the data in Experiment 2.

Experiment 3 contains the job file and the python used to generate the data in Experiment 3.

Note that both experiment 2 and experiment 3's job files generate other temp .sh files to job the individual jobs.

Experiment_real_data_2022 and Experiment_real_data_2023 within Experiment_4 are folders containing the files to replicate the results on the real datasets from Arizona. Run real_data_job.sh to replicate the results. Once the jobs are completed, the trace plots can be generated using the .sh file job_plot.sh

About

Code for Particle Markov Chain Monte Carlo.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published