Fig. 8
From: From pairs of most similar sequences to phylogenetic best matches
Performance comparison of two different outgroup choice methods: (1) outgroups chosen randomly from species in another branch of the root (green, same as in Fig. 7), (2) closest outgroups in all relative outgroup species corrected with Theorem 12 (red). Left and middle: Median (solid) and 10th percentile (dashed) of recall and precision as a function of noise level. Right: Boxplots of the F-measure differences for selected noise levels (method 2 − method 1)