Figure 4
From: Exact reconstruction of gene regulatory networks using compressive sensing
Guideline for designing new experiment based on coherence distribution. (Example 4) coherence distribution of the original Ω based on Exp#1 show that x 2 and x 4 cause high coherence. Thus, in order to reduce this high coherence, we should perturb either x 2 or x 4. We design two different experiment sets; for Exp#2, we perturb x 3 and for Exp#3, we perturb x 2. As we expected, Exp#3 is a more effective experiment to reduce the coherence between x 2 and x 4 and we recover the exact graph structure as shown in Figure 5. On the other hand, the coherence of Exp#2 remains almost the same as that of Exp#1 and we fail to recover the exact graph structure in Figure 5.