C-Command Software Forum

SS work with no corpus?

My corpus is a little large (5000) message and I am thinking about resetting it. However, I have SS set up to put the spam in the trash so creating a new corpus will require a significant amount of work.

Could I just reset the corpus and then just train it? Will that eventually build up the correct sized corpus?


Yes, but I recommend instead that you save up a few hundred spam messages first. That way the accuracy won’t drop in the interim between retting the corpus and re-training SpamSieve.

OK, I will switch SS to put the spam in a spam folder so I can track it.