BAS version 1.3.0 is being released for the launch of our Cousera Bayesian Statistics, part of the Statistics with R specialization. New features include a long awaited vignette with with an exampled carried out illustrating the functions, and major updates to the predict and fitted functions, such as identification of the “best predictive model”, the model that is closest to model averaging for prediction, standard deviations for fitted values and predictions, as well as credible intervals using the new confint function, based on S3 methods and plots for fitted and predicted values as well as coefficients.

data(Hald)
hald.ZS = bas.lm(Y ~ ., data=Hald, prior="ZS-null", modelprior=uniform())
plot(confint(coef(hald.ZS),parm=2:5))
plot(confint(predict(hald.ZS, se.fit=TRUE), parm="mean"))