Figure 6
From: IRIS: a method for reverse engineering of regulatory relations in gene networks
Comparison Kullback-Leibler divergence for (a) E. coli and (b) S. cerevisiae synthetic networks. The x-axis shows the biological noise levels used to generate the data sets. The y-axis represents the D KL values obtained as the mean for all tables of a network for the corresponding biological noise level. The Figure reports the D KL values of EM-MAP obtained using different discretisation approaches: IRIS (crosses), equal frequency (diamonds), global width (squares) and equal width (circles) and the D KL values obtained using IRIS algorithm both in discretisation step and in regulation function learning process (asterisks).