http://dsanalytics.com/dsblog/why-bayesian-spam-filters-are-doomed_92 Data Analytics- the art and science of analyzing data Sun, 06 Jul 2008 04:23:48 +0000 http://wordpress.org/?v=2.0.1 http://dsanalytics.com/dsblog/why-bayesian-spam-filters-are-doomed_92#comment-23 Thu, 22 Mar 2007 23:23:40 +0000 http://dsanalytics.com/dsblog/why-bayesian-spam-filters-are-doomed_92#comment-23 Bayesian approaches may not be the best algorithm for the job but there is also the issue of having enough training data... which is what I was alluding to in my gmail post. Systems like SpamAssassin also seem pretty interesting. Bayesian approaches may not be the best algorithm for the job but there is also the issue of having enough training data… which is what I was alluding to in my gmail post. Systems like SpamAssassin also seem pretty interesting.

]]> http://dsanalytics.com/dsblog/why-bayesian-spam-filters-are-doomed_92#comment-21 Thu, 22 Mar 2007 00:24:17 +0000 http://dsanalytics.com/dsblog/why-bayesian-spam-filters-are-doomed_92#comment-21 amazing how you and I can recognize spam in an instant, even the new and "improved" ones. But the algorithms get fooled all too often. Now there are a few emails I get that I'm not actually sure if they are spam or something I somehow signed up for once, but these are really the problem for me. But then again, there are many problems like this out there that we haven't found algorithms to crack--language translation and image classification are just two examples. amazing how you and I can recognize spam in an instant, even the new and “improved” ones. But the algorithms get fooled all too often. Now there are a few emails I get that I’m not actually sure if they are spam or something I somehow signed up for once, but these are really the problem for me. But then again, there are many problems like this out there that we haven’t found algorithms to crack–language translation and image classification are just two examples.

]]>