I'm trying to run a model comparison node in Enterprise Miner to compare the accuracy of two models, one of which was made in Miner, the other in R. I've tried to import a data source containing the actuals and predictions made by R into Enterprise Miner, then manually setting the roles to prediction and target but I'm getting an error message saying the model comparison node "must use one target variable modeled by a predecessor node". Is there an easy way to get SAS to treat the prediction from my R model as if it had been generated in SAS? Thanks.
... View more