Skip to main content


Springer Nature is making Coronavirus research free. View research | View latest news | Sign up for updates

Fig. 4 | EURASIP Journal on Bioinformatics and Systems Biology

Fig. 4

From: Using multi-step proposal distribution for improved MCMC convergence in Bayesian network structure learning

Fig. 4

Local maximum of a chain. An illustration of a local maximum for one of the five chains in Figs. 1, 2, and 3 during its sample phase. The nodes A to H represent different DAGs. Edges denote the possible transitions of length 1 between the DAGs, with the numbers indicating the acceptance probabilities (i.e., the probability of accepting the move if it was proposed). Possible transitions of length 2 are shown with dotted edges. One of the local maxima in the case of the one-step proposal consists of DAGs C and D, between which the chain oscillates, since transitions elsewhere are very improbable. Only edges representing transitions with acceptance probability higher than 0.0001 are shown to make the picture more readable

Back to article page