Gene Regulatory Networks from Multifactorial Perturbations Using Graphical Lasso: Application to the DREAM4 Challenge
Figure 2
Influence of the graphical Lasso penalty on network complexity and Bayesian Information.
A: Number of edges versus penalty for data set 3 in the multifactorial challenge with down arrows indicating the chosen associated with (from left to right) AIC, MAX_AUROC, MAX_AUPR and BIC. The horizontal line connects
and
of the 50 best BIC networks chosen in the ensemble network. B: BIC versus penalty for the five data sets.