Optimize attributes selection, preselect base attributes?

mafern76

If I want to have some attributes to be used as a starting subset for example when doing a forward selection, I see no way to do it in rapid miner other than joining them inside evaluation, which wouldn't be very efficient.

Is there a simpler way to do this or is there no other way at the moment?

Can be useful for example:

When you add new attributes.
When you add new cases.
When your pc crashes mid-process, using logs of course.

Thanks in advance.