Robustly fitting Gaussian graphical models: the R-package robFitConGraph

TL;DR

The paper introduces the R-package robFitConGraph, which offers a robust goodness-of-fit test for Gaussian graphical models, demonstrated through a music performance anxiety dataset, emphasizing the importance of robust model fitting.

Contribution

It provides a tutorial on using robFitConGraph, detailing its statistical foundation and practical application for robust Gaussian graphical model fitting.

Findings

Demonstrates robustness of the method on real data

Highlights advantages over traditional fitting methods

Provides practical guidance for users

Abstract

A tutorial-style introduction to the R-package robFitConGraph is given. The latter provides a robust goodness-of-fit test for Gaussian graphical models. Its use is demonstrated at a data example on music performance anxiety, which also illustrates why one would want to fit a Gaussian graphical model - and why one should do so robustly. The underlying statistical theory is briefly explained. The paper describes package version 0.4.1, available on CRAN from December 1, 2022. See https://CRAN.R-project.org/package=robFitConGraph