Skip to main content

Advertisement

Fig. 9 | EURASIP Journal on Bioinformatics and Systems Biology

Fig. 9

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

Fig. 9

Scores of MAP DAGs for different dataset sizes. Scores were calculated with the normal one-step proposal distribution (i.e., [p 1=1.0, p 2=0.0], marked p=1.0/0.0) and a proposal distribution with also steps of length 2 ([p 1=0.8,p 2=0.2], marked p=0.8/0.2). For each chain, the burn-in was 1,000,000 and the sample size was 50,000. The scores were calculated 10 times, using different sampled datasets

Back to article page