[SWLUG] Bayesian filters (Was: Reply to digest)

[Chris M. Jackson]

On Tue, 13 May 2003, [iso-8859-1] Rhys Sage wrote:

> As I understand it, Baysean filters combine three lists. List A is a list of
> words to ignore (these are grammatical words such as personal pronouns, conjunctions, possessive pronouns, demonstrative pronouns etc). List B is a list of the words found in spam mails. List C is words found in normal mails.
> 
> Incoming emails are checked against list C. If there're too many words found in the list, it's declared to be spam. (here, my understanding falters a bit).
Surely, you're thinking of list B, yes?  Either that, or your descriptions
of what's on each are muddled.

> Can anybody shed light on this interesting topic?

On the conditions that you stop sending HTML format mail to the list, and 
use meaningful subject lines, maybe.

-- 
|/  Chris M. Jackson
C>  Be seeing you!
H