Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Welcome to the python-psignifit 4 wiki!
This wiki hosts a manual and comments on our python clone of our MATLAB toolbox psignifit 4.
This code was tested against the MATLAB version and yields the same results up to numerical accuracy of the optimization algorithms for all fits and credible intervals (see here for details on our tests). However, this version is considerably less tested than the original MATLAB version.
Where to start?
First, install the toolbox.
Then, have a look at the Basic Usage or at the "demo_001.py" file, which cover similar content.
The wiki on the MATLAB version covers a broader discussion on how to apply our toolbox. Anything said there applies equally to the python version.
A paper describing our method in detail and showing tests for the congruency of our method is published at Vision Research: Painfree and accurate Bayesian estimation of psychometric functions for (potentially) overdispersed data by Heiko H. Schütt, Stefan Harmeling, Jakob H. Macke, and Felix A. Wichmann.
First, we would like to thank Sophie Laturnus and Ole Fortmann who programmed this python clone based on our MATLAB m-files.
Also, we would like to thank previous members of the Wichmann-lab who were involved in developing MCMC based methods of Bayesian inference for the psychometric function, most notably Ingo Fründ, Valentin Hänel, Frank Jäkel and Malte Kuss.
Furthermore, our thanks go to the reviewers of our manuscript and the students and colleagues who read the paper or tested the software and provided feedback: Nicole Eichert, Frank Jäkel, David Janssen, Britta Lewke, Lars Rothkegel, Joshua Solomon, Tom Wallis, Uli Wannek, Christian Wolf and Bei Xiao