This will only work for email accounts using IMAP.
You have this highlighted...
If I'm using a webmail client such as Roundcube, will it still work?
Just created the two folders and ran the commands. Everything worked fine and the results came out as expected. Now to put it in practice.Here we go guys, just posted: Training SpamAssassin's accuracy with sa-learn
Let's get some feedback.
It still needs to be in Seen so that sa-learn can learn "good" and "bad" tokens. However, after it's been placed in "Seen" and you know for a fact that sa-learn has been run (gaining tokens from the message), you can then delete said message(s) from Seen. - But, personally, I've never deleted an email except for spam =PAt what point would you delete email that you don't need? It may be good email, you just won't need it ever again. Still throw it in Seen?
Awesome, glad to hear it!Just created the two folders and ran the commands. Everything worked fine and the results came out as expected. Now to put it in practice.
Exactly. That's how I have mine setup.If I wanted to use a separate folders for particular emails, would I just run a ham-command for those folders as well?
@KH-DavidL Thanks for the followup. So far it's been great. Only one piece of spam has made it into my regular inbox since implementing the above process.
That may take a year..... or two.it seems that it's most accurate after reaching ~3,000 emails of ham and spam per email account
That may take a year..... or two.