Network/Node predictability
Date created: 2022-12-20
Predictability is a technique used in network analysis where each node is predicted by all the others. You then calculate the explained variance R2 for each node in the network. This is a Bayesian R2 because so far I have only seen it applied using the BGGM package which uses Bayesian Gaussian Graphical Models.
One measure of node centrality is predictability, i.e. how much of the variance of a node that is explained by all other nodes in the network. Predictability in the BGGM framework is estimated using Bayesian R2 with 95% credible intervals from the posterior distribution.
References
- Williams, D. R. (2021). Bayesian Estimation for Gaussian Graphical Models: Structure Learning, Predictability, and Network Comparisons. Multivariate Behavioral Research, 56(2), 336–352. https://doi.org/10.1080/00273171.2021.1894412