Skip to main content
. 2018 Mar 15;9:1090. doi: 10.1038/s41467-018-03424-4

Fig. 4.

Fig. 4

Influence of the number of samples on the performance of the top module detection methods. Shown are average training scores (left) and test scores (right) over all datasets and module definitions at different levels of random subsampling (five repeats)