Selecting a subste of attributes without choosing one by one

mansour

I am running various feature selection methods and for each method, a list of important features is created. I would like to generate new datasets based on these attributes. The only operator I see in Rapidminer is "Select Attribute" and then choosing Subset, but here I have to select the list (up to 350 attribute names manually). Is there any smart way to do this? I have at least 10 subsets and it is really difficult to do them manually.
Best regards.
Mansour

varunm1

Hello @mansour

You can connect to the output "fea" port of the feature selection method to "apply feature set" and also connect your dataset to the "exa" port to "apply feature set" operator. This will select the data (examples) related to attributes that were selected by your feature selection algorithm. This is applicable to automatic feature engineering operator and unsupervised feature selection operator as far as I know other operators like forward selection, backward elimination will give us the feature selected example set directly through their "exa" port. May I know which operators you are using?

Hope this helps. Please inform if you need more information.

varunm1

Hello @mansour

You can connect to the output "fea" port of the feature selection method to "apply feature set" and also connect your dataset to the "exa" port to "apply feature set" operator. This will select the data (examples) related to attributes that were selected by your feature selection algorithm. This is applicable to automatic feature engineering operator and unsupervised feature selection operator as far as I know other operators like forward selection, backward elimination will give us the feature selected example set directly through their "exa" port. May I know which operators you are using?

Hope this helps. Please inform if you need more information.