output posterior mean for beta for each component of prior mixture m,given data
comp_postmean(m, data)
m
mixture distribution with k components
data
details depend on the model