MiqG / learn2count

Structure learning based for zero-inflated negative binomial data

This package implements algorithms for structure learning of graphical models for count data.

For estimating a structure from given data, use function:

• pois.wald for Poisson node conditional models;
• nbscale.noT (or nb.wald using glm approach) for Negative Binomial node conditional models;
• zinb1.noT and zinb0.noT for zero-inflated negative binomial node conditional models.

For simulating a data set, see functions pois.simdata, nbinom.Simdata, and zinb.simdata for Poisson node conditional models, Negative Binomial node conditional models, and zero-inflated negative binomial node conditional models, respectively.

To install the package, you can run the following in R:

library(remotes)
install_github("drisso/learn2count")