SAS Enterprise Miner: Imbalance data

[sammmy]

Dear all,

I am working on a dataset where my target is having only 12% prevalence. that is giving me a high specificity and a very low sensitivity. Will it be fine to use SMOTE since the ratio is more than 10%? If not what other corrections can I use in Enterprise Miner to adjust for the imbalance. 

Thanks

[DougWielenga · Accepted solution]

For a discussions regarding SMOTE, please see 

https://communities.sas.com/t5/SAS-Data-Mining-and-Machine/SAS-Enterprise-Miner-SMOTE-sampling-with-...

https://communities.sas.com/t5/SAS-Data-Mining-and-Machine/SMOTE-with-missing-values/m-p/426917/high...

For a discussion of the issues with analyzing rare events, please see 

https://communities.sas.com/t5/SAS-Data-Mining-and-Machine/Help-with-over-under-sampling-of-the-rare...

For an approach to modeling rare events, consider looking at SAS note 47965 available at 

http://support.sas.com/kb/47/965.html

Hope this helps!

Cordially,

Doug