Evaluation metrics for incrementally growing inferred networks. Plots of TP/FP/FN/TN counts for networks inferred from different size datasets. All looks fairly monotonic and sensible. Good performance is obtained when learning from 1000+ samples. Poor performance on 100 or fewer samples.
Needham et al. BMC Systems Biology 2009 3:85 doi:10.1186/1752-0509-3-85