Skip to main content
Figure 10 | BMC Bioinformatics

Figure 10

From: Constructing non-stationary Dynamic Bayesian Networks with a flexible lag choosing mechanism

Figure 10

Comparison of reconstructed networks on drosophila data. A: The stationary directed networks predicted by [33]. B: The non-stationary undirected networks predicted by [34]. C: The non-stationary undirected networks predicted by RJnsDBNs with the setting of KNKT (λ s = 2) [1]. The posterior probability cutoff threshold 0.5 is used. D: The non-stationary undirected networks predicted by RJnsDBNs with the setting of UNUT (λ m = 2, λ s = 2). The posterior probability cutoff threshold 0.5 is used. The edges in blue color is the new edges different from G1. E: The non-stationary undirected networks predicted by FLnsDBNs (λ m = 35, λ s = 2). The posterior probability cutoff threshold 0.3 is used. The edges in blue color is the new edges different from the first graph G1.

Back to article page