Accounting for missing actors in interaction network inference from abundance data - Archive ouverte HAL Access content directly
Preprints, Working Papers, ... Year :

Accounting for missing actors in interaction network inference from abundance data

Abstract

Network inference aims at unraveling the dependency structure relating jointly observed variables. Graphical models provide a general framework to distinguish between marginal and conditional dependency. Unobserved variables (missing actors) may induce apparent conditional dependencies.In the context of count data, we introduce a mixture of Poisson log-normal distributions with tree-shaped graphical models, to recover the dependency structure, including missing actors. We design a variational EM algorithm and assess its performance on synthetic data. We demonstrate the ability of our approach to recover environmental drivers on two ecological datasets. The corresponding R package is available from github.com/Rmomal/nestor.

Dates and versions

hal-03134033 , version 1 (08-02-2021)

Identifiers

Cite

Raphaëlle Momal, Stephane S. Robin, Christophe Ambroise. Accounting for missing actors in interaction network inference from abundance data. 2021. ⟨hal-03134033⟩
53 View
0 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More