"Good" Corpus is too large. How do I reduce it?

As a new user, I apparently have made a new user mistake and trained too many good messages compared to the spam (112 good versus 52 spam). I still have the 52 spam messages in my “Spam” folder. Should I start over by clicking the corpus reset button then select the 52 current spam messages still in the Spam box and click “Train as Spam” to reset everything? Or is there a better way to get the good versus bad corpus back in line? (F.Y.I. I use Mac Mail.)

You should use the Reset Corpus command and then train it with the recommended ratio of messages.

Thanks for your help
Michael,

Thanks for your help. Your instructions took care of the problem.