Implementation of binary distribution in the Grassmann formalism, including conditional version. The Grassmann formalism was introduced in [1].
See the pdf file for more explanations.
- grassmann_distribution/:
- GrassmannDistribution: definition of Grassmann (gr) as well as mixture of Grassmann (mogr) distribution
- fit_grassmann: corresponding functions to estimate a gr / mogr (moment matching as well as MLE)
- conditional_grassmann: implements a conditional mogr in the same spirit as a MDN for a MoGauss
- notebooks/: some example notebooks how to define the distributions and an example to fit a mogr to dichotomized gauss data
- data: samples for a dichotomized gauss distribution, see [2] for details.
pip install GrassmannBinaryDistribution
[1] Arai, T. (2021). Multivariate binary probability distribution in the Grassmann formalism. Physical Review E, 103(6), 062104.
[2] Macke, J. H., et al. (2009). Generating spike trains with specified correlation coefficients. Neural computation 21.2: 397-423.