Sequential Monte Carlo Sampling for Abstract Data Types

Python classes describing sampling from distributions over abstract data types.

Includes Markov chain Monte Carlo (MCMC) sampling, Sequential Monte Carlo (SMC) sampling and decision forest (using sk-learn).

Published in DOI: XXXX

This software is licensed under Eclipse Public License 2.0. See LICENSE for more details.

This software is property of University of Liverpool and any requests for the use of the software for commercial use or other use outside of the Eclipse Public License should be made to University of Liverpool.

Copyright (c) 2023, University of Liverpool.

Requirements

The versions of Python and Python packages used for the results in XXX are as follows:

• Python 3.9.16
• numpy 1.19.2
• scipy 1.5.2
• scikit-learn 0.23.2
• pandas 1.1.3
• mpi4py 3.1.4

Either OpenMPI or Microsoft MPI should be installed to enable running with mpi4py for distributed parallel processing.

Running Tests

Tests can be found in examples with corresponding data sets.

Configuration of test cases is achieved by setting the following boolean variables (found in the example scripts):

MCMC_one_to_many = False
MCMC_many_to_one = False
SMC = False
DF = False

The number of particles/number of MCMC chains and iterations are also configured using the variables N and T respectively (also found in the example scripts).

The number of particles/number of MCMC chains and number of processor cores should be set to a power of 2 (e.g. 256, 512). The number of cores must be less than or equal to the number of particles, and is specified on the command line. Example scripts can be run using MPI with the following command:

mpiexec -n <NUM_CORES> python <path/to/script.py>