[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: SpamAssassin Corruption?
On 19 Feb 2004 Ken Hagan wrote:
> Perhaps adding a script that uses wordnet to determine if
> there are no sentences with verbs in the message? One could
> easily develop a decent grammar to check the grammar in the
> message and determine that there are no intelligible sentences
> and assume that the message is probably spam.
Given the example, I'd suggest a more elegant approach would
focus on statistical calculation of word length. Average word
length alone would determine the example text outta range for
ordinary English. I believe a certain absence of two and three
letter words would flag it at no expense to communicative text
in most any language using a Latin character set.
However, a test for intelligible language would be obstructive
to encrypted email which is unintelligible by design.
-
To unsubscribe, send email to majordomo@silug.org with
"unsubscribe silug-discuss" in the body.