A-NO-NE Music wrote: >I don't mind spam being missed. No big deal. I really do mind legit >message being missed and I fear if I accidentally delete them. >SpamAsassin never missed legit message. This is where I am at.
I think you can make SpamSieve a little less aggressive by using the Advanced > Spam-catching strategy slider. I had to move the slider a little to the left of middle, and had only two false negatives in the last few weeks. (At least I think so... ;-)). How big is your corpus, and what's the good/spam ratio in it? I guess the more even distributed it is, the better. Filling it with only Spam messages is probably counter-productive. Did you fill the corpus with a good batch of good messages from your database during setup? I selected a few hundred messages of my DB and added them as "good". My current corpus statistics are: Filtered Mail 13430 Good Messages 19572 Spam Messages (59%) 78 Spam Messages Per Day SpamSieve Accuracy 21 False Positives 265 False Negatives (93%) 99.1% Correct Corpus 12979 Good Messages 19194 Spam Messages (60%) 628040 Total Words Rules 6948 Blocklist Rules 1344 Whitelist Rules Showing Statistics Since 01.01.04 12:00 and it works quite well, at the moment. Regards, Christian.
