PLNmodels: Poisson lognormal models

The Poisson lognormal model and variants can be used for a variety of multivariate problems when count data are at play (including PCA for count data and network inference). This package implements an efficient algorithm to fit such models accompanied with a set of functions for vizualisation and diagnostic. Learn more in the vignettes: vignette(package = "PLNmodels").

Installation

System Requirements

Installation requires a system version of nlopt 2.4-2

• On Debian or Ubuntu use libnlopt-dev:

sudo apt-get install libnlopt-dev

• On Fedora or similar use NLopt-devel:

sudo yum install NLopt-devel

• With Mac OS X, install nlopt via homebrew

brew install nlopt

• On Windows, the package now builds and installs correctly, by including static libraries on compilation. However, I just test it with appveyor so I have never run PLNmodels on Windows: any feedbacks welcomed!

R Package installation

## w/o vignettes
devtools::install_github("jchiquet/PLNmodels")
devtools::install_github("jchiquet/PLNmodels", build_vignettes = TRUE)

References

Please cite our work using the following references:

• J. Chiquet, M. Mariadassou and S. Robin: Variational inference for probabilistic Poisson PCA, the Annals of Applied Statistics, to appear. link

• J. Chiquet, M. Mariadassou and S. Robin: Variational inference for sparse network reconstruction from count data, arXiv preprint, 2018. link