Fig. A.9.

This figure depicts the Precision-Recall curves and the corresponding AUPR values for each iteration in the networks used in this analysis. For many of the synthetic DREAM networks, the improvement in AUPR plateaus after 2 or 3 iterations, which may be due to the size of the networks. This may suggests that iterations improve RF, but larger networks are needed to verify this. The Precision-Recall curves themselves show minimal improvement after the second iteration for the empirical networks. However, the AUPR value increases as the number of iterations increases, even after the second iteration. Thus this shows that the addition of iterations to RF improves the AUPR values for the two empirical data sets used in this study.