I'm using the R package 'ranger' to make a regression random forest. The predictors (aka independent variables) include some interaction terms. When I ask the model to predict with some new data (exact same columns as training data), it complains: Error in predict.ranger.forest(forest, data, predict.all, num.trees, type, : Error: One or more independent variables not found in data.
Do I have to somehow create the interactions in the validation/prediction dataset? I don't know how. This doesn't make sense.
Example:
df <- data.frame( p1=1:10, p2=3:12, response=11:20)
train <- df[ 1:5, ]
valid <- df[ 6:10, ]
regr <- ranger( data = train, formula = response ~ p1 + p2 + p1*p2 )
pred <- predict( object=regr, data=valid )