Re: Spam problem

On 10/28/06, Scott G. Miller <[email protected]> wrote:
> It is possible, thats what I do for sisc-devel.
>
>         Scott

I'd like to take this opportunity to discuss the kind of spam the list
has been receiving. If you looked at them rather than deleting them,
you'd notice that they're basically legitimate but random words (not
porn/whatever related) in a nonsense order, probably randomly
generated with a simple trigram-type scheme. A lot of spam I get is
random words like that and then an impossible-to-OCR GIF, usually
about some penny stock or something.

This got through my email client's (Gmail, Google's attempt at web
mail) spam filter and I suspect it's because SpamAssassin heuristics
and naive Bayes are too weak for the job.

I think spam filter authors need to explore more sophisticated machine
learning architectures/algorithms, such as support vector machines.

Warren

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642