Hi all, I searched the archive, but was unable to find information on my questions. Here's hoping it wasn't due to poor search techniques.
My SA setup is for a single user Linux server. I have hand sorted 1768 ham messages and 785 spam messages received over the last year. Shortly after I began doing this I added a procmail recipe that sends anything with a score > 9 to /dev/null. I'm not using auto-learn. 1) Will the absence of high scoring messages from my spam corpus adversely affect the results of manual training? 2) I've not included in my ham corpus any messages from lists, as I have rules in .procmail that deliver them directly without sending them through SA. Does this seem like a good strategy? Thanks for any suggestions or feedback! Ted -- Theodore Heise <[EMAIL PROTECTED]> Bloomington, IN, USA ------------------------------------------------------- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01 _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
