Tony Finch wrote:
> On Wed, 24 Mar 2004, Nigel Wade wrote:
>
>>Ian A B Eiloart wrote:
>>
>>>But Baysian filters have their limitations. Researchers already have lists
>>>of words that these filters tend to like, and techniques for refining those
>>>lists on a per filter basis.
>>
>>Do you have any references for that?
>>The only reports I've read have been for a very specific set of words.
>
>
> One of the things that doesn't impress me about SpamBayes' and
> SpamAssassin's statistical filter (they use the same basic design) is that
> it works on the basis of presence of words, not frequency of words. This
> makes them relatively easy to fool. I don't know if DSPAM is cleverer or
> not -- it isn't usable here because it doesn't work with Cyrus.
>
I've seen no evidence yet of the Bayesian filter in SA being particularly
easy to fool. If it was I'd expect to be seing a lot more spam in my inbox.
Of the last 3500+ messages I've received I've had 4 false negatives and no
false positives.
--
Nigel Wade, System Administrator, Space Plasma Physics Group,
University of Leicester, Leicester, LE1 7RH, UK
E-mail : nmw@???
Phone : +44 (0)116 2523548, Fax : +44 (0)116 2523555
