Skip to content
Model fitting using Approximate Bayesian Computation
Branch: master
Clone or download
Mike Irvine
Latest commit 04bddc6 Jun 6, 2018
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
ABCPRC
scripts
LICENSE
README.md
Tutorial_Ecology.ipynb New Tutorial. Heavy refactoring of module to include 2D data. Feb 20, 2018
Tutorial_Epidemiology.ipynb
Tutorial_SIS.ipynb
setup.py
single_distribution_example.py Added parralelism, unit test framework and others. Apr 1, 2016

README.md

ABCPRC

ABCPRC is an Approximate Bayesian Computation Particle Rejection Scheme designed to perform model fitting on individual-based models.

Setup

To setup, first download a local copy and then run

python setup.py install

Introduction

Import as

import ABCPRC as prc

A fitting class is setup using

m = prc.ABC()

You can then either use the built-in tolerances or fit your own using

m.fit()

The fitting can then be performed using

m.run(num_particles)

and the results shown (using seaborn),

m.trace(plot=True)

Tutorials

Three example tutorials accompany this package.

Testing

Tests are run using the python nose2 package. To install run

pip install nose2
pip install cov-core

and tests can be performed running the command

nose2 --with-coverage
You can’t perform that action at this time.