ABCPRC is an Approximate Bayesian Computation Particle Rejection Scheme designed to perform model fitting on individual-based models.
To setup, first download a local copy and then run
python setup.py install
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)
Three example tutorials accompany this package.
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