cran / bart Goto Github PK

Is it currently possible to use a cluster with mc.wbart instead of using a single computer with multiple cores?

The vignette doesn't really provide an example of the use of the random effects feature. It's not completely clear how to incorporate the grouping variable. There is the u.train specification, but what if the same groups are available for the predictions? There is no "u.test". Should the grouping variable be part of x.test? Or does the function assume that the random effects are new "draws" from the same distribution? Thanks.