All RevoScaleR functions use SAS contrasts by default, instead of the default R treatment contrasts. To get the same results from both functions you have two options: i) Before running your code, set the following global options: options(contrasts=c(factor="contr.SAS",ordered="contr.poly")) ii) In your call to lm() or glm, add the option 'contrasts=list(Species=contr.SAS)'. Include all factor variables in this statement.  Â
QA: Why does rxLinMod produce different results than lm?
Applies To
Revolution Analytics