On Mon, Apr 15, 2013 at 2:19 AM, Pratik Sarkar <iampratiksarkar@gmail.com>wrote:

Yes,thats why the language model concept will come in handy. The classifier will be trained on the existing archives. For eg,in debian list,the classifier will be trained on a training set containing copyright law,source code,etc.Whereas in an economists list,the classifier will be trained on the training set containing economics-related emails and conversations. Whenever an economics-related message will try to enter the debian list,the classifier would check the language model of the (economics)message with the classifier and would classify it correctly as the sentence construction/language model of a technical message(debian list message) and an economics message varies a lot. Please feel free to as me,if I have not been able to explain my views. Regards Pratik

On Mon, Apr 15, 2013 at 2:07 AM, Paul Wise <pabs3@bonedaddy.net> wrote:

...

On Mon, Apr 15, 2013 at 4:55 PM, Stephen J. Turnbull wrote:

...

That's a great idea (and a great example of why I don't think spam-filtering technology should be embedded in a Mailman handler! :-)

I think you might not be thinking expansively enough, it allows more flexibility to do spam filtering in multiple places. Different lists will have different definitions of what is an unwanted message. On the Debian lists we don't want anything about shares, liquidity and financial instruments, but some economists will find messages about source code and copyright law unwanted on their lists. A list for pharmacists will probably want to accept messages containing usually-spammy keywords. A generic mailing-list-hoster hosting all 3 of those will want to allow most mail and add per-list filters.

-- bye, pabs

http://bonedaddy.net/pabs3/

---

Mailman-Developers mailing list Mailman-Developers@python.org http://mail.python.org/mailman/listinfo/mailman-developers Mailman FAQ: http://wiki.list.org/x/AgA3 Searchable Archives: http://www.mail-archive.com/mailman-developers%40python.org/ Unsubscribe: http://mail.python.org/mailman/options/mailman-developers/iampratiksarkar%40...

Security Policy: http://wiki.list.org/x/QIA9