ALL FEATURE REQUESTS HERE ARE MONITORED BY OUR PRODUCT TEAM.
VOTING MATTERS!
IDEAS WITH HIGH NUMBERS OF VOTES (USUALLY ≥ 10) ARE PRIORITIZED IN OUR ROADMAP.
NOTE: IF YOU WISH TO SUGGEST A NEW FEATURE, PLEASE POST A NEW QUESTION AND TAG AS "FEATURE REQUEST". THANK YOU.
VOTING MATTERS!
IDEAS WITH HIGH NUMBERS OF VOTES (USUALLY ≥ 10) ARE PRIORITIZED IN OUR ROADMAP.
NOTE: IF YOU WISH TO SUGGEST A NEW FEATURE, PLEASE POST A NEW QUESTION AND TAG AS "FEATURE REQUEST". THANK YOU.
The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here
Automatic feature engineering: including and excluding features
kypexin
RapidMiner Certified Analyst, Member Posts: 291 
Unicorn
Hi there,
Is there a simple way to explicitly exclude certain features from optimization / selection process? Which means, I want to run feature selection, but I have a few features that I want to leave in a feature set no matter what, so they are not automatically removed during selection.
UPDATE: I have actually found a workaround, however it complicates the process flow a bit: using SET ROLE to change role of given attributes from regular to smth else, and then change it back to regular before training final model and applying it to test subset:
This also requires that APPLY FEATURE SET keeps originals and makes special role for them:
So far, my concern is whether it would be more convenient just to have an additional setting in AUTOMATIC FEATURE ENGINEERING operator to explicitly exclude or include a certain attributes subset?
Is there a simple way to explicitly exclude certain features from optimization / selection process? Which means, I want to run feature selection, but I have a few features that I want to leave in a feature set no matter what, so they are not automatically removed during selection.
UPDATE: I have actually found a workaround, however it complicates the process flow a bit: using SET ROLE to change role of given attributes from regular to smth else, and then change it back to regular before training final model and applying it to test subset:
This also requires that APPLY FEATURE SET keeps originals and makes special role for them:
So far, my concern is whether it would be more convenient just to have an additional setting in AUTOMATIC FEATURE ENGINEERING operator to explicitly exclude or include a certain attributes subset?
Tagged:
5
Comments
I think this can be implemented for example using standard operator setting for choosing attributes (like in SELECT ATTRIBUTES): single, subset, etc
Vladimir
http://whatthefraud.wtf