20
$\begingroup$

I am looking for an implementation of the linear ballistic accumulator model or Ratcliff's diffusion model (e.g. in R, MATLAB, or Python).

$\endgroup$
0

5 Answers 5

17
$\begingroup$

Here are a few options. I have not tried them yet personally.

As mentioned below, the rtdists package in R is able to fit both LBA and diffusion models.

Other Options

LBA

Scott Brown has a copy of Donkin et al (2009) on his web page with some code in R, Excel, and WinBUGS for fitting the LBA model:

There's also the glba package on CRAN by Ingmar Visser.

Diffusion model

The Diffussion model is available as a matlab toolbox called (DMAT).

References

  • Donkin, C., Averell, L., Brown, S.D., & Heathcote, A. (2009) Getting more from accuracy and response time data: Methods for fitting the Linear Ballistic Accumulator model. Behavior Research Methods, 41, 1095-1110. PDF
  • Vandekerckhove, J., & Tuerlinckx, F. (2008). Diffusion model analysis with MATLAB: A DMAT primer. Behavior Research Methods, 40, 61-72. doi:10.3758/BRM.40.1.61 PDF
$\endgroup$
0
9
$\begingroup$

For the diffusion model, there is also Eric-Jan Wagenmakers' "EZ-diffusion model", which you can find here.

This paper compares three different pieces of software for estimation of diffusion model parameters:

von Ravenzwaaij D., & Oberauer, K. (2009). How to use the diffusion model: Parameter recovery of three methods: EZ, fast-dm, and DMAT. Journal of Mathematical Psychology, 53 (6), 463–473. [PDF]

$\endgroup$
5
$\begingroup$

The R package diffIRT (http://www.dylanmolenaar.nl/jss1265.pdf) estimates both the Q and the D diffusion models (see his website for the van der Maas et al. paper discussing the differences between these models). R code for the EZ2 approach, which is much faster if that is important for your applications, is http://raoul.socsci.uva.nl/EZ2/.

$\endgroup$
3
$\begingroup$

In the subheading you also mention that you're interested in matlab / python implementations:

I've personally used DMAT in matlab at that's a nice package. However, the python based HDDM package may be one of the best around at the moment (in my opinion) and it has a good user guide.

http://ski.clps.brown.edu/hddm_docs/abstract.html

and the paper associated with the package:

Wieki et al (2013): http://journal.frontiersin.org/article/10.3389/fninf.2013.00014/full

References

$\endgroup$
2
$\begingroup$

The R package rtdists is another great option:

Provides response time distributions (density/PDF, distribution function/CDF, quantile function, and random generation): (a) Ratcliff diffusion model based on C code by Andreas and Jochen Voss and (b) linear ballistic accumulator (LBA) with different distributions underlying the drift rate.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.