BGGM: A R Package for Bayesian Gaussian Graphical Models

Dr, Williams; Mulder, Joris

doi:10.31234/osf.io/3b5hf

Cited by 26 publications

(23 citation statements)

References 45 publications

(60 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Using the replication code provided by Brandt (2020), we first used a Bayesian Gaussian graphical model with the BGGM package (Williams & Mulder, 2019) to estimate the psychological network among 19 the political belief items. 4 These items were selected to serve as the nodes in the belief network in the original study because they "were available across the three waves and...were measures of political attitudes" (Brandt, 2020, p. 3).…”

Section: Methodsmentioning

confidence: 99%

Out of bounds? The boundary specification problem for centrality in psychological networks

Neal¹,

Neal²

2020

Preprint

View full text Add to dashboard Cite

Analysis of psychological networks has become common in multiple subfields including clinical, social, and personality psychology, where the focus is often on identifying highly central nodes. However, the boundary of these networks is often ambiguous and relevant nodes are often not represented in the network. This can limit the replicability of centrality measures. In this paper, we demonstrate that the rank order of nodes by centrality measures in a political belief network cannot be replicated in networks that are identical except that they have narrower (i.e., fewer nodes) or wider (i.e., more nodes) boundaries. We conclude by recommending that centrality measures should be computed and interpreted only in psychological networks for which a theoretically meaningful boundary can be specified and (nearly) all the nodes inside that boundary are included in the network.

show abstract

Section: Methodsmentioning

confidence: 99%

Out of bounds? The boundary specification problem for centrality in psychological networks

Neal¹,

Neal²

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The 1 -regularized networks were estimated with the R package qgraph using the default settings, in addition to bootnet for bootstrapping the estimated networks (Epskamp, Borsboom, & Fried, 2018). The Bayesian models were fitted with the package BGGM (Williams & Mulder, 2019b).…”

Section: Illustrative Examplementioning

confidence: 99%

Bayesian Estimation for Gaussian Graphical Models: Structure Learning, Predictability, and Network Comparisons

Williams

2021

Multivariate Behavioral Research

View full text Add to dashboard Cite

Gaussian graphical models (GGM; "networks") allow for estimating conditional independence structures that are encoded by partial correlations. This is accomplished by identifying nonzero relations in the inverse of the covariance matrix. In psychology the default estimation method uses 1 -regularization, where the accompanying inferences are restricted to frequentist objectives. Bayesian methods remain relatively uncommon in practice and methodological literatures. To date, they have not yet been used for estimation and inference in the psychological network literature. In this work, I introduce Bayesian methodology that is specifically designed for the most common psychological applications. The graphical structure is determined with posterior probabilities, which allow for assessing conditional dependent and independent relations. Additional methods are provided for extending inference to specific aspects withinand between-networks, including partial correlation differences and Bayesian methodology to quantify network predictability. I first demonstrate that the decision rule based on posterior probabilities can be calibrated to the desired level of specificity. The proposed techniques are then demonstrated in several illustrative examples. The methods have been implemented in the R package BGGM.

show abstract

“…We study 13 different estimation algorithms, further described in Table 1. In particular, we study the EBICglasso and ggmModSelect (two variants) algorithms that are implemented in the qgraph R package; Epskamp, Cramer, Waldorp, Schmittmann, & Borsboom, 2012), two variants of full information maximum likelihood (FIML) and weight least squares (WLS) estimation as in the psychonetrics R package (Epskamp, 2020a(Epskamp, , 2020b, two variants of mixed graphical model estimation as implemented in the mgm R package (Haslbeck & Waldorp, 2015), two variants of Bayesian estimation as implemented in the BGGM R package (Williams & Mulder, 2019), and two more unregularized estimation procedures as implemented in the GGMnonreg R package .…”

Section: Aim Of the Papermentioning

confidence: 99%

Which Estimation Method to Choose in Network Psychometrics? Deriving Guidelines for Applied Researchers

Isvoranu¹,

Epskamp²

2021

Preprint

View full text Add to dashboard Cite

The Gaussian Graphical Model (GGM) has recently grown popular in psychological research, with a large body of estimation methods being proposed and discussed across various fields of study, and several algorithms being identified and recommend as applicable to psychological datasets. Such high-dimensional model estimation, however, is not trivial, and algorithms tend to perform differently in different settings. In addition, psychological research poses unique challenges, including placing a strong focus on weak edges (e.g., bridge edges), handling data measured on ordered scales, and relatively limited sample sizes. As a result, there is currently no consensus regarding which estimation procedure performs best in which setting. In this large-scale simulation study, we aimed to overcome this gap in the literature by comparing the performance of several estimation algorithms suitable for gaussian and skewed ordered categorical data across a multitude of settings, as to arrive at concrete guidelines from applied researchers. In total, we investigated 60 different metrics across 564,000 simulated datasets. We summarized our findings through a platform that allows for manually exploring simulation results. Overall, we found that an exchange between discovery (e.g., sensitivity, edge weight correlation) and caution (e.g., specificity, precision) should always be expected and achieving both¬—which is a requirement for perfect replicability—is difficult. Further, we identified that the estimation method is best chosen in light of each research question and highlighted, alongside desirable asymptotic properties and low sample size discovery, results according to most common research questions in the field.

show abstract

BGGM: A R Package for Bayesian Gaussian Graphical Models

Cited by 26 publications

References 45 publications

Out of bounds? The boundary specification problem for centrality in psychological networks

Out of bounds? The boundary specification problem for centrality in psychological networks

Bayesian Estimation for Gaussian Graphical Models: Structure Learning, Predictability, and Network Comparisons

Which Estimation Method to Choose in Network Psychometrics? Deriving Guidelines for Applied Researchers

Contact Info

Product

Resources

About