C-Command Software Forum

Will not train as good mail

For the last year I have been clicking 2 particular emails as good and every day the next update goes directly to the spam folder. It’s the same type of email for both each day, just 2 local newspaper headlines that I constantly try to train as good and it never works. Any idea how to fix this?

Thanx In Advance!!!

Maybe these messages aren’t going to the spam folder because of SpamSieve. Have you checked the log to see whether it says that SpamSieve is predicting these messages to be spam?

It appears that the keywords were marked numerous times as both good and spam, so I removed them all and I will see tomorrow if I have to revise any of them.

Thanx…I was unaware I could do that as the last time I looked years ago I didn’t see how it could be done an so I reset the entire corpus

It’s important to determine whether SpamSieve predicted the message to be good or spam or neither. This is because no amount of training will help if SpamSieve isn’t able to see all the incoming messages.

Although it is possible to edit individual words it the corpus, this is a lot of work and probably not very effective because you don’t know all the words that were in the message. If, instead, you correct the mistakes using the training commands, SpamSieve will make sure that the words have the proper counts associated with them.

So far it looks like that fixed it…