Hello, I'm starting to use the combination bnlearn and Rgraphviz, and it generates wonderful bayesian network plots! Congratulations! However I'm facing some strange issue: bayesian networks generated from some specific dataframe have text inside ellipses very big, and it's unformatting the plot. The commands I used are as follow: # load data frame d <- read.csv2("financ_auto.csv", row.names=1) # omit NA's d <- na.omit( d ) # Iterate all variables for(i in 1:ncol(d)) { if(is.numeric(d[[i]])) { # Categorize variable if it's numerical d[[i]] <- cut(d[[i]],5) } } # Generate bn object (graph) bn <- mmpc(d) # Generate pdf tf <- tempfile() pdf( tf, width = 20, height = 20) graphviz.plot(bn, shape="ellipse", layout="neato") dev.off() tf The dataframe file and generated plot are on the following links: http://www.mediafire.com/file/zp58zg6l28hn49m/financ_auto.csv http://www.mediafire.com/file/l4lj4gzkohw8yhy/plot.pdf Does anyone knows what may be happening? Thanks in advance! Cheers, Joao D. Duarte UFMG - Brasil [[alternative HTML version deleted]]