Hello, I have been using dspam for a while now and its been very good.
I have recently noticed tho that a lot of the emails that I get from mailing lists are being marked as spam by dspam, this is OK as my procmail filters keep them out of my junk folder but I just get a little worried about legit email getting marked as spam by dspam. Like an email I just got from amazon for a order confirmation got marked as spam. I think maybe one of the reasons for this is that dspam is scanning message headers and learning them as spam? or am I just talking gobldygoop? Also a setting I have always used is to set Train on error, I did this because I didn't want to mess up dspam's training by not training emails that I may have missed that dspam got wrong (I get just under 100 emails an hour that dspam thinks is spam and I just dont have time to go through every one). Would it be better to train on everything in my case ? I wouldn't mind relaxing dspams filtering a bit as I also use public databases to help detect spam, it this easily possible? Thanks for any help and advice, Matt. -- Matt Richards
signature.asc
Description: Digital signature
