On Thu, 7 Mar 2002, Matthew Cline wrote:
> Alright, here's a first pass at it:
>
> score DOT_HIDING 1.0
> score DOT_HIDING_3 2.0
> score DOT_HIDING_5 2.0
I think I'd score DOT_HIDING itself as 0.0. The chances you could conceal
a spam from all the other rules by hiding just one word are pretty small.
Maybe more like
score DOT_HIDING 0.0
score DOT_HIDING_3 1.5
score DOT_HIDING_5 2.5
> # Grab all words with a dot in them
> my @suspects = map(/(\S*\w\.\w\S*)/g, @{$body});
Couldn't other punctuation be used just as easily as dots?
> # Get rid of URIs
> # Get rid of email adresses
> # Get rid of acronyms
It might also be good to filter out honorifics \b(Mrs?|Ms|Dr).\w+ (there
are probably more that could be included).
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
