https://communities.sas.com/t5/SAS-Data-Science/Difference-between-variable-importance-and-feature-engineering/m-p/636100#M8198 <P>Hi Team,</P><P>&nbsp;</P><P>I was interested in getting mutual information for variables in a churn model. I see that under feature engineering we have PCA,SVD, auto encoder etc.. I would like to know what could be the best information criteria to use instead of mutual information</P> Tue, 31 Mar 2020 06:42:03 GMT prajnat 2020-03-31T06:42:03Z https://communities.sas.com/t5/SAS-Data-Science/Difference-between-variable-importance-and-feature-engineering/m-p/636100#M8198 <P>Hi Team,</P><P>&nbsp;</P><P>I was interested in getting mutual information for variables in a churn model. I see that under feature engineering we have PCA,SVD, auto encoder etc.. I would like to know what could be the best information criteria to use instead of mutual information</P> Tue, 31 Mar 2020 06:42:03 GMT https://communities.sas.com/t5/SAS-Data-Science/Difference-between-variable-importance-and-feature-engineering/m-p/636100#M8198 prajnat 2020-03-31T06:42:03Z https://communities.sas.com/t5/SAS-Data-Science/Difference-between-variable-importance-and-feature-engineering/m-p/636157#M8199 They are the SAME thing I think . <BR />PCA is Primary Component Analysis.<BR />SVD is a decompose way of matrix.<BR /><BR />I would like to call <BR /><a href="https://communities.sas.com/t5/user/viewprofilepage/user-id/13684">@Rick_SAS</a> Tue, 31 Mar 2020 10:40:31 GMT https://communities.sas.com/t5/SAS-Data-Science/Difference-between-variable-importance-and-feature-engineering/m-p/636157#M8199 Ksharp 2020-03-31T10:40:31Z