Is bogofilter Bayesian?

[Boris 'pi' Piwinger]

Greg Louis wrote:

>> Very interesting here is what Liudvikas Bukys added about
>> AdaBoost
>> (http://kiew.cs.uni-dortmund.de:8001/mlnet/instances/81d91e8d-dc15ed23e9)
>> which maybe gives a theoretical background for what I found
>> just by trying and from very unscientific ideas.
> 
> This wouldn't be worth replying to (it was discussed several months
> ago) but that people have short memories. 

I must have missed that.

> Adaboost and related
> techniques will very effectively train for recognition of a member of a
> population, at the expense of capability to recognize something
> _similar_ to the members of a population. 

Maybe you have something here where to point at for further
reading. Maybe you can add a comment to Gary's page.

> Whether a particular
> training method proves effective or overspecific needs to be determined
> empirically.

Right. Have you seen an implementation of Adaboost for spam
filters?

pi