Hi, >> While some of the mail from that sender seems legitimate, other mail >> clearly isn't, but it has the same header as a legitimate mail, making >> it very difficult to properly train bayes or otherwise accurately >> determine that it's indeed spam and it should be discarded. > > I wouldn't obsess over it. Bayes is pretty good at picking out the relevant > markers of messages and ignoring irrelevant parts. Train the spam as spam > and the non-spam as non-spam and Bayes should eventually figure it out. > > [That's my experience, at any rate. However, we use our own Bayes > implementation that works a little differently from the built-in SA version, > so maybe SA will behave differently...]
Yes, thanks, good advice. The difficulty is determining which of it bayes should learn as ham and which as spam, of course. With at least this sender, the delineation is very subjective. Do people really subscribe to this crap? :-) Thanks, Alex
