From: Fritz B. <fb...@iw...> - 2006-04-24 16:08:15
|
>To my knowledge, there is nothing wrong with or less effective about >ASSP's Bayesian processing. It can be argued, that it is one of the best. People, who are not familiar with ASSP tend to compare in the wrong way. ASSP is using a method of building a "trusted" net. So, it is the combination of whitelist and bayesian score, which makes it effective. With 1000 mails in the corpus, you can make no judgement. |