I created Time Series model and got a bunch of derived features.
I don't want to generate derived features in Time Series model.
Is there an option to avoid generating derived features?
Solved! Go to Solution.
The exact option name is Exclude features from derivation. Sorry for the confusion.
I cannot find "do not derive" under the advanced option.
It's possible to set all the features as "do not derive" under the advanced options section on the UI before the project starts. Alternatively, the option default_to_do_not_derive can be set to True, if you use the python client.
import datatrobot as dr spec = dr.DatetimePartitioningSpecification(default_to_do_not_derive=True)
Anyway, DataRobot will use the first lag for all numerical features except the target to avoid leakage.