Fig. 16
From: Constructing phylogenetic networks via cherry picking and machine learning
Comparison of the results of FeatImp, ML and TrivialRand on the instance groups described in Sects. "Experiments on normal data", "Experiments on LGT data" and "Experiments on real data". Each point on the horizontal axis corresponds to an instance group; each line represents the average, within the instance group, of the output reticulation number divided by the reference value. The shaded areas represent 95% confidence intervals