Over the last few weeks or months (I’ve lost track), I’ve been getting loads of spam about stocks and movie production. I train each one as spam, and SS corrects itself. But they keep getting through. They use many of the same words, so I’d think that SS would have caught on by now. One explanation is that whoever’s behind these is just busier than ever. But my total spam count is about the same or just a little bit more.
I’ve checked the settings, and it’s all normal. SS does filter out most other spam (except now and then when it seems to fail completely and let through a bunch of even the most obvious ones).
Thanks for sending the additional information. It looks like there’s too much (and possibly old and/or incorrect) data in SpamSieve’s corpus. I suggest that you choose Reset Corpus from the Filter menu, then re-train SpamSieve with a smaller number of recent messages.
Really? The log showed more than 10K messages in SpamSieve’s corpus. If you do an initial training with fewer than 1,000 messages and thereafter only train SpamSieve with mistakes, it would normally take many years to get to that level. Do you receive many messages per day?
At least 200 a day. There are already 111 in the spam bucket since last night at 11, and I’m sure it’ll double or even triple by tonight. At that rate, I’ll get to 10K in 45 days. Is that a lot?
I don’t consider that a lot. Most of the messages SpamSieve will classify correctly, so they won’t be trained, and they won’t contribute to the 10K. You can also turn off auto-training to slow down the growth of the corpus.