Skip to main content
. 2016 Jan 22;6:410. doi: 10.3389/fphys.2015.00410

Figure 1.

Figure 1

Similarity index of the models built by the different algorithms. The Jaccard index was computed for each pair of models and clustered in function of the Euclidean distance. Contrary to what was expected, the output models of the tested algorithms, despite having been fed with the same input show a huge variability.The descritization-based algorithms (GIMME, iMAT, Akesson, FASTCORE, and FASTCORMICS) show the highest similarity levels.