The current version of SAS EM does not support the use of a weight variable within the decision tree node. It would be great if there is a weight statement for the decision tree node. Just like many other standard procedures, e.g. weighted ordinary least squares.
Regression trees that can incorporate a weight variable are very useful for purposes of cost prediction. In many cases, individuals may not be enrolled for the entire contract period/year. In order to obtain accurate cost predictions, it is necassary to weigh the target variable (=costs) for the fraction the individual was enrolled. So, it would be very useful if there is a weight statement for the decision tree node.