[OT]Re: [oclug]Port 4662
bart-oclug at jukie.net
Sun Jan 26 07:22:40 EST 2003
* Dave O'Neill <dmo at acm.org> [030126 03:05]:
> On Sat, Jan 25, 2003 at 11:25:52PM -0500, Bart Trojanowski wrote:
> > * Dave O'Neill <dmo at acm.org> [030125 09:12]:
> > > Technical solutions to nontechnical problems are doomed to failure.
> > Dave, can you not train your Bayesian mail-filter to mark off-topic
> > posts automatically?
> Well, you've entirely missed the point here... there is no feasable software
> solution to filter out off-topic posts. Sure, I can set up a filtering
> rule (or train bogofilter to the point where [OT] flags something as spam),
> but it relies on wetware to do the work of marking each and every off-topic
> post as [OT].
You missed the point. If it is possible for a statistical function to
predict if something is a SPAM (off-topic under all contexts), it is
possible to predict if it is off-topic in the OCLUG context... with
> Besides which, I've given up on my bogofilter experiment, as I've found
> that rule-based filtering tends to do a much better job of keeping spam away
> from my inbox.
You should try Razor, it's statistical based, but the training set is
provided by everyone, not just you. Hence there is a much higher chance
of successful matches then if you were to train it yourself.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 189 bytes
Desc: not available
Url : http://tux.oclug.on.ca/pipermail/oclug/attachments/20030126/4c3fde8d/attachment.bin
More information about the OCLUG