>In the release notes you mention that HEUR has been set to have mutliple
>levels. How does this benefit us? Does a level 1 return a weight of 1 and
>so on?
You're a step ahead of the class. <G>
The weighting now gives new life to the heuristics test. Where it wasn't
useful before as a test by itself, it can become very useful with the
weighting. Before, it would calculate the likelihood of spam, and if a
certain threshold was reached, it would be considered spam. However, it's
now been broken down into 10 different levels. Every E-mail will be
assigned a value between 1 (almost certainly not spam) to 10 (almost
certainly spam).
You can now define the heuristics test like this:
HEUR heuristics [level] x [weight] 0
Here, "level" is the 1-10 value of how likely the E-mail is to be spam, and
"weight" is the weight assigned if the E-mail is at that level.
For example, you could have "HEUR10 heuristics 10 x 8 0". This would
define a heuristics test named HEUR10, that would get set off if the
likelihood of spam was at the level 10 (almost certainly spam), and would
have a weight of 8.
Another example would be "HEUR1 heuristics 1 x -5 0". This would define a
heuristics test named HEUR1, that would get set off if the likelihood of
spam was at the level 1 (almost certainly NOT spam), and would have a
weight of -5. This means that if the heuristics test is pretty sure that
the E-mail is not spam, it could help prevent a false positive that would
otherwise occur.
-Scott
---
This E-mail came from the Declude.JunkMail mailing list. To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail". You can E-mail
[EMAIL PROTECTED] for assistance. You can visit our web
site at http://www.declude.com .
