Skip to main content
Fig. 2 | EURASIP Journal on Bioinformatics and Systems Biology

Fig. 2

From: Graph reconstruction using covariance-based methods

Fig. 2

Simulation study for network deconvolution (ND). a Illustration of a graph with direct and indirect edges of different order; original graph is a chain graph. b Simulations conducted on the graph depicted in (a) with different scaling parameters. Shown is the log discriminative ratio given in (29). c Effect of network deconvolution on direct and indirect edges. If the indirect edges are clearly separable in the covariance matrix, then network deconvolution can better separate them from direct edges. d If direct and indirect edges are not separable in the covariance matrix, then network deconvolution cannot separate them too

Back to article page