skip to navigation
skip to content

addm-toolbox 0.1.8

A toolbox for data analysis using the attentional drift-diffusion model.

Latest Version: 0.1.12

This toolbox can be used to perform model fitting and to generate simulations for the attentional drift-diffusion model (aDDM), as well as for the classic version of the drift-diffusion model (DDM) without an attentional component.

Prerequisites

aDDM-Toolbox supports Python 2.7 only and requires the following libraries: * deap * matplotlib * numpy * pandas * scipy

Installing

$ pip install addm_toolbox

Running tests

To make sure everything is working correctly after installation, try (from a UNIX shell, not the Python interpreter):

$ addm_run_tests

This should take a while to finish, so maybe go get a cup of tea :)

Getting started

To get a feel for how the algorithm works, try:

$ addm_demo --display-figures

You can see all the arguments available for the demo using:

$ addm_demo --help

Here is a list of useful scripts which can be similarly run from a UNIX shell: * addm_demo * ddm_pta_test * addm_pta_test * addm_pta_mle * addm_pta_map * addm_simulate_true_distributions * addm_basinhopping * addm_genetic_algorithm * ddm_mla * addm_mla

You can also have a look directly at the code in the following modules: * addm.py contains the aDDM implementation, with functions to generate model simulations and obtain the likelihood for a given data trial. * ddm.py is equivalent to addm.py but for the DDM. * addm_pta_test.py generates an artificial data set for a given set of aDDM parameters and attempts to recover these parameters through maximum a posteriori estimation. * ddm_pta_test.py is equivalent to addm_pta_test.py but for the DDM. * addm_pta_mle.py fits the aDDM to a data set by performing maximum likelihood estimation. * addm_pta_map.py performs model comparison for the aDDM by obtaining a posterior distribution over a set of models. * simulate_addm_true_distributions.py generates aDDM simulations using empirical data for the fixations.

Common issues

Make sure you are using the toolbox under Python 2.7, not Python 3.

If you get a Python RuntimeError with the message “Python is not installed as a framework.”, try creating the file ~/.matplotlib/matplotlibrc and adding the following code:

backend: TkAgg

Authors

License

This project is licensed under the GNU GENERAL PUBLIC LICENSE - see the COPYING file for details.

Acknowledgments

This toolbox was developed as part of a research project in the Rangel Neuroeconomics Lab at the California Institute of Technology.

 
File Type Py Version Uploaded on Size
addm_toolbox-0.1.8-py2-none-any.whl (md5) Python Wheel py2 2017-10-12 1017KB
addm_toolbox-0.1.8.tar.gz (md5) Source 2017-10-12 992KB